Trial In Primary Biliary Cirrhosis Research Articles

Random forest and support vector machine based hybrid liver disease detection

This study develops an automated liver disease detection system using a support vector machine and random forest detection techniques. These techniques are trained on data containing the information collected from the Mayo Clinic trial in primary biliary cirrhosis (PBC) of the liver conducted between 1974 and 1984. The proposed system can detect the presence of liver disease in the test set. The random forest model is used for recursive feature elimination at the pre-processing stage and the support vector machine is trained on the optimal feature set. The experimental result shows that the proposed support vector machine (SVM) model has achieved 78.3% accuracy.

Statistical inference for proportional mean residual life model with continuous auxiliary covariate

In many statistical studies involving failure data, the main covariate is only measured in a validation set due to financial and technical restriction, while some auxiliary information for the main covariate is collected for the full cohort. Making use of the auxiliary information will increase the efficiency of the study. In this paper, we consider statistical inference for the proportional mean residual life model when the primary covariate is measured only for a randomly chosen subcohort and some auxiliary variables are available for the whole study cohort. To further make use of the auxiliary information to improve the efficiency of the study, we propose anestimating equation based on the inverse probability of censoring weighting techniques. The resulting estimators are shown to be consistent and asymptotically normal. Extensive simulations are conducted to evaluate the finite sample performance of the proposed methods. We illustrate the proposed method with a real data set from the Mayo Clinic trial in primary biliary cirrhosis of the liver.

Weighted Kaplan-Meier estimators motivating to estimate HIV-1 RNA reduction censored by a limit of detection.

Measuring the magnitude of reduction in HIV-1 RNA levels accurately is difficult because many patients have a censored reduction due to the limit of detection (LOD) of the virologic assay being employed. The use of censored methods has improved the analysis of such reductions compared with crude methods but implies independent censoring. For HIV-1 RNA reduction data, the value at which a patient's HIV-1 RNA reduction becomes censored is mainly determined by the patient's baseline HIV-1 RNA level. We suggest two possibilities based on modification of the redistribution to the right algorithm to handle the situation of dependence either from a single continuous marker, that is, the baseline HIV-1 RNA level, or from multiple markers. Two series of simulation, one in the HIV-1 RNA setting and one in the classical censoring setting, compared performance of the previous methods with our suggestions. Our proposed estimators show good performances when the dependent censoring is due to LOD. Overall, in the classical censoring setting, our suggestions perform as well as other methods including the Inverse Probability of Censoring Weighted and the Kaplan-Meier imputation with Bootstrap. We applied those estimators to estimate the HIV-1 RNA reduction at week 8 of 502 patients who received a raltegravir-containing regimen and to data from the Mayo Clinic trial in primary biliary cirrhosis.

SURVIVAL MODELS IN THE CONTEXT OF MULTIPLICATIVE AND ADDITIVE HAZARDS

In survival analysis, multiplicative and additive hazards models provide the two principal frameworks to study the association between the hazard and covariates. When these models are considered for analyzing a given survival dataset, it becomes relevant to evaluate the overall goodness-of-fit and how well each model can predict those subjects who subsequently will or will not experience the event. In this paper, this evaluation is based on a graphical representation of the Cox-Snell residuals and also on a time-dependent version of the area under the receiver operating characteristic (ROC) curve, denoted by AUC(t). A simulation study is carried out to evaluate the performance of the AUC(t) as a tool for comparing the predictive accuracy of survival models. A dataset from the Mayo Clinic trial in primary biliary cirrhosis (PBC) of the liver is also considered to illustrate the usefulness of these tools to compare survival models formulated under distinct hazards frameworks.

New treatments/targets for primary biliary cholangitis.

SummaryPrimary biliary cholangitis (PBC) is an autoimmune, cholestatic, chronic liver disease that ultimately progresses towards cirrhosis and liver failure if untreated. While ursodeoxycholic acid has been established as standard of care for PBC in the last few decades, significant advances in second-line treatment options have recently been made and new therapeutic developments are currently under evaluation. The purpose of this article is to provide the clinician with an overview of the current treatment options and future opportunities for patients with PBC.

Semiparametric Bayesian analysis of accelerated failure time models with cluster structures.

In this paper, we develop a Bayesian semiparametric accelerated failure time model for survival data with cluster structures. Our model allows distributional heterogeneity across clusters and accommodates their relationships through a density ratio approach. Moreover, a nonparametric mixture of Dirichlet processes prior is placed on the baseline distribution to yield full distributional flexibility. We illustrate through simulations that our model can greatly improve estimation accuracy by effectively pooling information from multiple clusters, while taking into account the heterogeneity in their random error distributions. We also demonstrate the implementation of our method using analysis of Mayo Clinic Trial in Primary Biliary Cirrhosis.

Time-dependent ROC curve analysis in medical research: current methods and applications

BackgroundROC (receiver operating characteristic) curve analysis is well established for assessing how well a marker is capable of discriminating between individuals who experience disease onset and individuals who do not. The classical (standard) approach of ROC curve analysis considers event (disease) status and marker value for an individual as fixed over time, however in practice, both the disease status and marker value change over time. Individuals who are disease-free earlier may develop the disease later due to longer study follow-up, and also their marker value may change from baseline during follow-up. Thus, an ROC curve as a function of time is more appropriate. However, many researchers still use the standard ROC curve approach to determine the marker capability ignoring the time dependency of the disease status or the marker.MethodsWe comprehensively review currently proposed methodologies of time-dependent ROC curves which use single or longitudinal marker measurements, aiming to provide clarity in each methodology, identify software tools to carry out such analysis in practice and illustrate several applications of the methodology. We have also extended some methods to incorporate a longitudinal marker and illustrated the methodologies using a sequential dataset from the Mayo Clinic trial in primary biliary cirrhosis (PBC) of the liver.ResultsFrom our methodological review, we have identified 18 estimation methods of time-dependent ROC curve analyses for censored event times and three other methods can only deal with non-censored event times. Despite the considerable numbers of estimation methods, applications of the methodology in clinical studies are still lacking.ConclusionsThe value of time-dependent ROC curve methods has been re-established. We have illustrated the methods in practice using currently available software and made some recommendations for future research.

Estimating the Treatment Effect in a Clinical Trial Using Difference in Restricted Mean Survival Time

The causal effect of a new medical treatment compared with a standard regimen is best assessed in a randomized controlled trial setting. When the main outcome is time to some event of interest, such as death, studies often use the hazard ratio to describe the treatment effect. Typically, proportional hazards are assumed. Here I discuss several significant disadvantages of using the hazard ratio, including its vulnerability to the proportionality assumption, its relative nature, and its lack of relationship with time-to-event or survival probabilities. I describe the use of restricted mean survival time as an alternative outcome measure in time-to-event trials. With this method, the treatment effect is defined as the difference in restricted mean between the trial arms. I suggest the use of Royston and Parmar's (2002, Statistics in Medicine 21: 2175–2197) class of flexible parametric models, implemented through the command stpm2 (Lambert and Royston [2009, Stata Journal 9: 265–290] and Andersson and Lambert [2012, Stata Journal 12: 623–638]), to estimate the required quantities. With this approach, proportional hazards are not assumed. I describe a new command, strmst, for implementing these calculations. This method supports “direct” adjustment for covariates by using marginalization over their observed distribution, and it supports estimation of treatment effects conditional on fixed values of covariates. I illustrate the methodology using data from a trial in primary biliary cirrhosis. I provide an example that demonstrates the importance of understanding the relationship between the treatment effect, the prognosis of the disease outcome, and the often-neglected time domain.

Buckley-James estimator of AFT models with auxiliary covariates.

In this paper we study the Buckley-James estimator of accelerated failure time models with auxiliary covariates. Instead of postulating distributional assumptions on the auxiliary covariates, we use a local polynomial approximation method to accommodate them into the Buckley-James estimating equations. The regression parameters are obtained iteratively by minimizing a consecutive distance of the estimates. Asymptotic properties of the proposed estimator are investigated. Simulation studies show that the efficiency gain of using auxiliary information is remarkable when compared to just using the validation sample. The method is applied to the PBC data from the Mayo Clinic trial in primary biliary cirrhosis as an illustration.

Estimation of a monotone percentile residual life function under random censorship

In this paper, we introduce a new estimator of a percentile residual life function with censored data under a monotonicity constraint. Specifically, it is assumed that the percentile residual life is a decreasing function. This assumption is useful when estimating the percentile residual life of units, which degenerate with age. We establish a law of the iterated logarithm for the proposed estimator, and its n-equivalence to the unrestricted estimator. The asymptotic normal distribution of the estimator and its strong approximation to a Gaussian process are also established. We investigate the finite sample performance of the monotone estimator in an extensive simulation study. Finally, data from a clinical trial in primary biliary cirrhosis of the liver are analyzed with the proposed methods. One of the conclusions of our work is that the restricted estimator may be much more efficient than the unrestricted one.

Accelerated Failure Time Models with Auxiliary Covariates

In this paper we study semi-parametric inference procedure for accelerated failure time models with auxiliary information about a main exposure variable. We use a kernel smoothing method to introduce the auxiliary covariate to the likelihood function. The regression parameters are then estimated through maximization of the estimated likelihood function. A consistent estimator of the variance of the estimator of the regression coefficients is proposed. Simulation studies show that the efficiency gain is remarkable when compared to just using the validation sample. The method is applied to the PBC data from the Mayo Clinic trial in primary biliary cirrhosis as an illustration.

Optimizing biochemical markers as endpoints for clinical trials in primary biliary cirrhosis

Biochemical tests have been recommended as endpoints for clinical trials in primary biliary cirrhosis (PBC) because the use of liver transplantation and death as endpoints in ursodeoxycholic acid (UDCA) therapeutic trials is unfeasible. The best inclusion criteria cut-off values and cut-off for demonstrating treatment success have not been defined. Our aim was to determine the optimal biochemical values for patient inclusion and to define values for treatment success in therapeutic trials. We performed a retrospective review of 73 patients with PBC treated with UDCA followed over 36months. Following one year of UDCA therapy, the likelihood of developing clinical endpoints of varices, ascites, encephalopathy, death or transplantation over the ensuing two years, based on degrees of elevation of biochemical markers, was analyzed using chi-square or Fisher's exact test. Patients with ALP≥2 X upper limit of normal (ULN) had a 2-fold greater likelihood of developing endpoints compared to patients with lower values (23% versus 11%), (p<0.05). Patients with bilirubin>1mg/dL were 4 times more likely to develop endpoints compared to those with lower values (33% versus 8%), (p=0.02). These values help identify the patient population for adjunctive therapy trials. Patients with ALP ≤1.67 X ULN and bilirubin ≤1mg/dL demonstrated the least likelihood of reaching adverse clinical endpoints and can be used to define treatment success. Optimal ALP and Bilirubin levels can be used as appropriate biochemical criteria for patient selection and defining treatment success in future clinical trials in patients with PBC.

A stunning year for advances in biliary tract disease

For the last several years, I have had the privilege of editing the biliary section for Current Opinion in Gastroenterology. However, this has been my favorite year. The year has shed light on many clinical and scientific problems in biliary tract pathobiology. I am excited to share these discoveries with you, as highlighted in the following paragraphs. Perhaps the most significant advance was the publication by Bridgewater et al. (pp. 258–261) establishing a new platform of care for patients with advanced biliary tract cancer, namely administration of cisplatin and gemcitabene. For the first time in medical oncology, we now have a standard of care for these cancers. The authors not only highlight this study but also point out that the data cannot be extrapolated to the adjuvant setting. Hopefully, the success of this study will spur additional clinical trials in this disease. Baron (pp. 262–267) describes a unique experience with removable, covered metallic stents for benign biliary tract disease. Heretofore, metallic stents were only used for malignant biliary strictures because they were ‘permanent’. With the advent of these removable covered stents, we now have new armamentarium to attack recalcitrant and problematic benign biliary strictures. The author’s success with these stents is described in detail. We are beginning to realize that cholangiocytes have an ‘activated phenotype’ that contributes to liver disease progression, especially biliary fibrosis. Furthermore, this phenotype is retained by cholangiocarcinoma cells explaining the highly desmoplastic nature of these cancers. Omenetti and Diehl (pp. 268–275) in their pioneering work suggest that Hedgehog and platelet-derived growth factor signaling pathways contribute to the activated cholangiocyte phenotype and associated periductular fibrosis. The fibrosis is due to cross talk between activated cholangiocytes and myofibroblasts. Al Sirica (pp. 276–284) explores similar cross talk between cholangiocarcinoma cells and cancer-associated fibroblasts as a key process in biliary tract cancer progression. These mediators are potential therapeutic targets for the treatment of benign and malignant biliary tract disease. Although ursodeoxycholic acid is effective for the treatment of primary biliary cirrhosis (PBC), not all patients respond. Thus, there is an unmet need for new therapy. Following identification of the farnesoid X receptor (FXR) nuclear receptor in biliary physiology, studies in humans employing an FXR agonist have been long awaited. Lindor (pp. 285–288) summarizes the role of FXR agonists in liver physiology relevant to the biliary tract and the initial results of a clinical trial in PBC employing the FXR agonist obeticholic. The results are promising, displaying a significant reduction in serum cholestatic biomarkers. These biomarkers, such as the serum alkaline phosphatase, have been verified as indicators of disease activity in PBC. We await further trials, but FXR-directed therapy with obeticholic appears to be disease modifying in this liver disease population. All clinicians are aware that cholestasis is associated with pruritus. However, the nature of the pruritus mediator has remained elusive. Elferink et al. (pp. 289–293) provides compelling data that lysophosphatidic acid (LPA) is the serum mediator of pruritus. A serum enzyme autotoxin is responsible for the generation of LPA from lysophosphatidyl choline. This is as an extremely exciting observation. We now have a molecule that we can measure and, hopefully, therapeutically modulate to relieve pruritus in patients with cholestasis. Polycystic liver disease and polycystic liver kidney disease are unique forms of biliary tract disease characterized by numerable hepatic cysts lined by biliary epithelium. The cysts progressively enlarge over time, causing mass-like symptoms in a subset of patients. The genotypes for these diseases have been defined, but an established medical therapy is unavailable. Drenth et al. (pp. 294–300) describe the use of somatostatin analogues for these patients, which appear to be successful by decreasing biliary epithelial cell secretion into these cysts. The failure of mammalian target of rapamycin inhibitors to modify disease progression is also reviewed. Finally, an emerging controversy has arisen regarding the risk of recurrent primary sclerosing cholangitis (PSC) in patients undergoing liver donor transplantation for this disease. The reported data from Japan suggest a high rate for recurrent PSC. Graziadei (pp. 301–305) reviews the existing literature and notes the limitations of the data. Clearly, there is yet insufficient information to alter clinical practice, and PSC patients should continue to be acceptable candidates for live donor liver transplantation.

Optimizing Biochemical Markers as Endpoints for Clinical Trials in Primary Biliary Cirrhosis

Background Biochemical tests have been recommended as endpoints for clinical trials in primary biliary cirrhosis (PBC) because the use of liver transplantation and death as endpoints in ursodeoxycholic acid (UDCA) therapeutic trials is unfeasible. The best inclusion criteria cut-off values and cut-off for demonstrating treatment success have not been defined. Aim Our aim was to determine the optimal biochemical values for patient inclusion and to define values for treatment success in therapeutic trials. Methods We performed a retrospective review of 73 patients with PBC treated with UDCA followed over 36 months. Following one year of UDCA therapy, the likelihood of developing clinical endpoints of varices, ascites, encephalopathy, death or transplantation over the ensuing two years, based on degrees of elevation of biochemical markers, was analyzed using chi-square or Fisher's exact test. Results Patients with ALP≥2 X upper limit of normal (ULN) had a 2-fold greater likelihood of developing endpoints compared to patients with lower values (23% versus 11%), (p 1 mg/dL were 4 times more likely to develop endpoints compared to those with lower values (33% versus 8%), (p = 0.02). These values help identify the patient population for adjunctive therapy trials. Patients with ALP ≤1.67 X ULN and bilirubin ≤1mg/dL demonstrated the least likelihood of reaching adverse clinical endpoints and can be used to define treatment success. Conclusion Optimal ALP and Bilirubin levels can be used as appropriate biochemical criteria for patient selection and defining treatment success in future clinical trials in patients with PBC.

American Association for the Study of Liver Diseases Endpoints Conference: Design and Endpoints for Clinical Trials in Primary Biliary Cirrhosis

American Association for the Study of Liver Diseases Endpoints Conference: Design and Endpoints for Clinical Trials in Primary Biliary Cirrhosis

Targeting Farnesoid X Receptor in Hepatic and Biliary Inflammatory Diseases

Targeting Farnesoid X Receptor in Hepatic and Biliary Inflammatory Diseases

Reflections on therapeutic trials in primary biliary cirrhosis: A quality of life oriented counter-view

To the editor: A recent placebo-controlled, multi-center clinical trial, published in Hepatology, has reported that the addition of methotrexate to ursodeoxycholic acid (UDCA) therapy has no discernable effect on survival, or time to transplantation, in primary biliary cirrhosis (PBC).1 Although obviously disheartened by the efforts he and his colleagues have put in to what transpired to be a negative trial, we feel that the accompanying, rather pessimistic, editorial by Combes is far too narrow in its outlook.2 In this editorial Combes argues that it is difficult, if not impossible, to envisage future improvement in therapy in PBC. Whilst morbidity and mortality are of fundamental importance as goals in the treatment of any disease, we would argue that it is just as important that clinicians look to treat patients' symptoms and to improve their quality of life (QOL). This is particularly important in a condition such as PBC where significant impairment of QOL is recognised, typically resulting from symptoms such as fatigue, which are unrelated to disease severity.3-6 The importance of improvement in QOL as a goal of therapy in its own right, is made all the more pertinent by the observation, made so elegantly by Combes and colleagues, that outcomes in terms of survival are actually very good, almost regardless of therapy in young PBC patients with early disease (although it should be noted that the outcome in terms of survival may be less rosy in older patient groups).7 Whilst the Combes trial looked at end points including mortality, liver transplantation and clinical parameters, there is no mention of symptomatic and/or QOL assessment. In light of the low mortality demonstrated, and the likelihood that, if the patients are at all representative, they have substantial (but non-assessed) impairment of their QOL, it could be argued that the most important outcome measure was not studied. The counter view to Combes is that what is needed in PBC is not pessimism but realism about the true nature of the impact that the disease has on patients and, accordingly, a paradigm shift in treatment goals. Although the search for novel additional pharmacological therapies able to stop disease progression continues, there is a persuasive argument that patients who progress on UDCA would do so with other agents, and that transplantation will be required by those patients who do progress in any case. However, even in patients in whom the disease is only slowly progressive, the symptoms of the disease can have a major impact on QOL. Ameliorating symptoms and improving QOL in such patients would be of major practical benefit to large numbers of patients. We have a duty of care as clinicians and researchers to look at QOL in our PBC patients and we would suggest that, in the future, at least as much effort be put in to clinical research and trial work in the area of improving QOL, as in to high concept, large scale and long term mortality-orientated trials which as Combes acknowledges have generated relatively little benefit to patients to date. David E.J. Jones*, Neeraj Bhala*, Julia L. Newton*, * Liver Research Group, School of Clinical Medical Sciences, University of Newcastle, Newcastle-upon-Tyne, UK

Reflections on Therapeutic Trials in Primary Biliary Cirrhosis * #

Reflections on Therapeutic Trials in Primary Biliary Cirrhosis * #

Dichotomizing continuous predictors in multiple regression: a bad idea

In medical research, continuous variables are often converted into categorical variables by grouping values into two or more categories. We consider in detail issues pertaining to creating just two groups, a common approach in clinical research. We argue that the simplicity achieved is gained at a cost; dichotomization may create rather than avoid problems, notably a considerable loss of power and residual confounding. In addition, the use of a data-derived 'optimal' cutpoint leads to serious bias. We illustrate the impact of dichotomization of continuous predictor variables using as a detailed case study a randomized trial in primary biliary cirrhosis. Dichotomization of continuous data is unnecessary for statistical analysis and in particular should not be applied to explanatory variables in regression models.

Quasi-likelihood estimation for relative risk regression models

For a prospective randomized clinical trial with two groups, the relative risk can be used as a measure of treatment effect and is directly interpretable as the ratio of success probabilities in the new treatment group versus the placebo group. For a prospective study with many covariates and a binary outcome (success or failure), relative risk regression may be of interest. If we model the log of the success probability as a linear function of covariates, the regression coefficients are log-relative risks. However, using such a log-linear model with a Bernoulli likelihood can lead to convergence problems in the Newton-Raphson algorithm. This is likely to occur when the success probabilities are close to one. A constrained likelihood method proposed by Wacholder (1986, American Journal of Epidemiology 123, 174-184), also has convergence problems. We propose a quasi-likelihood method of moments technique in which we naively assume the Bernoulli outcome is Poisson, with the mean (success probability) following a log-linear model. We use the Poisson maximum likelihood equations to estimate the regression coefficients without constraints. Using method of moment ideas, one can show that the estimates using the Poisson likelihood will be consistent and asymptotically normal. We apply these methods to a double-blinded randomized trial in primary biliary cirrhosis of the liver (Markus et al., 1989, New England Journal of Medicine 320, 1709-1713).