The landscape of etiological patterns of hepatocellular carcinoma and intrahepatic cholangiocarcinoma in Thailand.

Thailand is among countries with the highest global incidence and mortality rates of hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA). While viral hepatitis and liver fluke infections have been associated with HCC and iCCA, respectively, other environmental risk factors, overall risk factor commonality and combinatorial roles, and effects on survival have not been systematically examined. We conducted a TIGER-LC consortium-based population study covering all high-incidence areas of both malignancies across Thailand: 837 HCC, 1474 iCCA, and 1112 controls (2011-2019) were comprehensively queried on lifelong environmental exposures, lifestyle, and medical history. Multivariate logistic regression and Cox proportional hazards analyses were used to evaluate risk factors and associated survival patterns. Our models identified shared risk factors between HCC and iCCA, such as viral hepatitis infection, liver fluke infection, and diabetes, including novel and shared associations of agricultural pesticide exposure (OR range of 1.50; 95% CI: 1.06-2.11 to 2.91; 95% CI: 1.82-4.63) along with vulnerable sources of drinking water. Most patients had multiple risk factors, magnifying their risk considerably. Patients with lower risk levels had better survival in both HCC (HR 0.78; 95% CI: 0.64-0.96) and iCCA (HR 0.84; 95% CI: 0.70-0.99). Risk factor co-exposures and their common associations with HCC and iCCA in Thailand emphasize the importance for future prevention and control measures, especially in its large agricultural sector. The observed mortality patterns suggest ways to stratify patients for anticipated survivorship and develop plans to support medical care of longer-term survivors, including behavioral changes to reduce exposures.

A latent functional approach for modeling the effects of multidimensional exposures on disease risk.

Understanding the relationships between exposure and disease incidence is an important problem in environmental epidemiology. Typically, a large number of these exposures are measured, and it is found either that a few exposures transmit risk or that each exposure transmits a small amount of risk, but, taken together, these may pose a substantial disease risk. Further, these exposure effects can be nonlinear. We develop a latent functional approach, which assumes that the individual effect of each exposure can be characterized as one of a series of unobserved functions, where the number of latent functions is less than or equal to the number of exposures. We propose Bayesian methodology to fit models with a large number of exposures and show that existing Bayesian group LASSO approaches are a special case of the proposed model. An efficient Markov chain Monte Carlo sampling algorithm is developed for carrying out Bayesian inference. The deviance information criterion is used to choose an appropriate number of nonlinear latent functions. We demonstrate the good properties of the approach using simulation studies. Further, we show that complex exposure relationships can be represented with only a few latent functional curves. The proposed methodology is illustrated with an analysis of the effect of cumulative pesticide exposure on cancer risk in a large cohort of farmers.

Uncovering circadian rhythms in metabolic longitudinal data: A Bayesian latent class modeling approach.

Researchers in biology and medicine have increasingly focused on characterizing circadian rhythms and their potential impact on disease. Understanding circadian variation in metabolomics, the study of chemical processes involving metabolites may provide insight into important aspects of biological mechanism. Of scientific importance is developing a statistical rigorous approach for characterizing different types of 24-hour patterns among high dimensional longitudinal metabolites. We develop a latent class approach to incorporate variation in 24-hour patterns across metabolites where profiles are modeled with finite mixtures of distinct shape-invariant circadian curves that themselves incorporate variation in amplitude and phase across metabolites. An efficient Markov chain Monte Carlo sampling is used to carry out Bayesian posterior computation. When the model was fit separately by individual to the data from a small group of participants, two distinct 24-hour rhythms were identified, with one being sinusoidal and the other being more complex with multiple peaks. Interestingly, the latent pattern associated with circadian variation (simple sinusoidal curve) had a similar phase across the three participants, while the more complex latent pattern reflecting diurnal variation differed across individual. The results suggested that this modeling framework can be used to separate 24-hour rhythms into an endogenous circadian and one or more exogenous diurnal patterns in describing human metabolism.

IFN-λ4 is associated with increased risk and earlier occurrence of several common infections in African children.

Genetic polymorphisms within the IFNL3/IFNL4 genomic region, which encodes type III interferons, have been strongly associated with clearance of hepatitis C virus. We hypothesized that type III interferons might be important for the immune response to other pathogens as well. In a cohort of 914 Malian children, we genotyped functional variants IFNL4-rs368234815, IFNL4-rs117648444, and IFNL3-rs4803217 and analyzed episodes of malaria, gastrointestinal, and respiratory infections recorded at 30,626 clinic visits from birth up to 5 years of age. Compared to children with the rs368234815-TT/TT genotype (IFN-λ4-Null), rs368234815-dG allele was most strongly associated with an earlier time-to-first episode of gastrointestinal infections (p = 0.003). The risk of experiencing an infection episode during the follow-up was also significantly increased with rs368234815-dG allele, with OR = 1.53, 95%CI (1.13-2.07), p = 0.005 for gastrointestinal infections and OR = 1.30, 95%CI (1.02-1.65), p = 0.033 for malaria. All the associations for the moderately linked rs4803217 (r2 = 0.78 in this set) were weaker and lost significance after adjusting for rs368234815. We also analyzed all outcomes in relation to IFN-λ4-P70S groups. Our results implicate IFN-λ4 and not IFN-λ3 as the primary functional cause of genetic associations with increased overall risk and younger age at first clinical episodes but not with recurrence or intensity of several common pediatric infections.

Bayesian network meta-regression hierarchical models using heavy-tailed multivariate random effects with covariate-dependent variances.

Network meta-analysis (NMA) is gaining popularity in evidence synthesis and network meta-regression allows us to incorporate potentially important covariates into network meta-analysis. In this article, we propose a Bayesian network meta-regression hierarchical model and assume a general multivariate t distribution for the random treatment effects. The multivariate t distribution is desired for heavy-tailed random effects and converges to the multivariate normal distribution when the degrees of freedom go to infinity. Moreover, in NMA, some treatments are compared only in a single study. To overcome such sparsity, we propose a log-linear regression model for the variances of the random effects and incorporate aggregate covariates into modeling the variance components. We develop a Markov chain Monte Carlo sampling algorithm to sample from the posterior distribution via the collapsed Gibbs technique. We further use the deviance information criterion and the logarithm of the pseudo-marginal likelihood for model comparison. A simulation study is conducted and a detailed analysis from our motivating case study is carried out to further demonstrate the proposed methodology.

Bayesian inference for network meta-regression using multivariate random effects with applications to cholesterol lowering drugs.

Low-density lipoprotein cholesterol (LDL-C) has been identified as a causative factor for atherosclerosis and related coronary heart disease, and as the main target for cholesterol- and lipid-lowering therapy. Statin drugs inhibit cholesterol synthesis in the liver and are typically the first line of therapy to lower elevated levels of LDL-C. On the other hand, a different drug, Ezetimibe, inhibits the absorption of cholesterol by the small intestine and provides a different mechanism of action. Many clinical trials have been carried out on safety and efficacy evaluation of cholesterol lowering drugs. To synthesize the results from different clinical trials, we examine treatment level (aggregate) network meta-data from 29 double-blind, randomized, active, or placebo-controlled statins +/$-$ Ezetimibe clinical trials on adult treatment-naïve patients with primary hypercholesterolemia. In this article, we propose a new approach to carry out Bayesian inference for arm-based network meta-regression. Specifically, we develop a new strategy of grouping the variances of random effects, in which we first formulate possible sets of the groups of the treatments based on their clinical mechanisms of action and then use Bayesian model comparison criteria to select the best set of groups. The proposed approach is especially useful when some treatment arms are involved in only a single trial. In addition, a Markov chain Monte Carlo sampling algorithm is developed to carry out the posterior computations. In particular, the correlation matrix is generated from its full conditional distribution via partial correlations. The proposed methodology is further applied to analyze the network meta-data from 29 trials with 11 treatment arms.

A contemporary amniotic fluid volume chart for the United States: The NICHD Fetal Growth Studies-Singletons.

BACKGROUND: Amniotic fluid is essential to normal fetal development and is estimated clinically with ultrasound scanning to identify pregnancies that are at risk for poor perinatal outcome. OBJECTIVE: Our goal was to develop a United States standard for amniotic fluid volume that is estimated by the amniotic fluid index and single deepest pocket. STUDY DESIGN: We performed a planned secondary analysis of a multicenter observational study of 2334 low-risk women with normal singleton gestations from 1 of 4 self-reported racial/ethnic groups. Eligible women had confirmed first-trimester dating criteria with health status, lifestyles, and medical and obstetric histories that were associated with normal fetal growth. Consenting women underwent serial (up to 5) sonographic evaluations of amniotic fluid between 15 and 40 weeks of gestation after being assigned randomly to 1 of 4 gestational age observation schedules. Twelve United States perinatal centers participated, and all sonograms were performed by credentialed sonographers who used identical, high-resolution equipment; caregivers were unaware of results but were notified for oligohydramnios. Women (n=597) who were subsequently found to have clinically significant antepartum complications were excluded. Racial/ethnic-specific nomograms for amniotic fluid index and single deepest pocket across gestation were developed with the use of linear mixed models with cubic splines; racial/ethnic differences were evaluated both with global and between-group tests. Median, 3rd, 5th, 10th, 90th, 95th and 97th percentile values were also estimated. We further considered the possible confounding effects of selected maternal characteristics and the estimated fetal weight at each sonogram. RESULTS: A total of 1719 pregnant women met inclusion criteria and had available data. These included 480 non-Hispanic white women, 418 non-Hispanic black women, 485 Hispanic women, and 336 Asian women. Both the amniotic fluid index and the single deepest pocket varied across gestation with maximal values at 26 and 33 weeks of gestation, respectively. Statistically significant differences were observed by maternal race/ethnicity. The between-group differences that were observed at 17-22 and 35-40 weeks of gestation remained statistically significant after adjustment for maternal characteristics and estimated fetal weight. These between-group racial/ethnic differences were most prominent after 35 weeks of gestation and at the extremes of dispersion (3rd and 97th percentiles). All 3rd and 97th percentile amniotic fluid index values were within the range of commonly used cutoffs to define oligohydramnios (≤5 cm) and polyhydramnios (≥25 cm). However, the 3rd percentile values ranged between 5.9 cm at 40 weeks of gestation and 10.1 cm at 25-27 weeks of gestation; the 97th percentile values ranged between 24.8 cm at 38 weeks of gestation and 15.7 cm at 15 weeks of gestation. CONCLUSION: Sonographic amniotic fluid volume estimates vary by racial/ethnic group, but the absolute differences appear to be small and may not be clinically significant. Selected maternal characteristics and estimated fetal weight did not affect the racial/ethnic differences. Between-group differences are maximal after 35 weeks of gestation and at the extremes of the upper and lower dispersion estimates. Given the observed variability in extreme (3rd and 97th percentile) dispersion values over the gestation, use of single cutoffs to define out-of-range measurements may not be appropriate clinically. These data might form a contemporary United States standard for amniotic fluid estimation that uses the amniotic fluid index and the single deepest pocket.

Fetal growth patterns in pregnancy-associated hypertensive disorders: NICHD Fetal Growth Studies.

BACKGROUND: Fetal growth patterns in pregnancy-associated hypertensive disorders is poorly understood because prospective longitudinal data are lacking. OBJECTIVE: The objective of the study was to compare longitudinal fetal growth trajectories between normotensive women and those with pregnancy-associated hypertensive disorders. STUDY DESIGN: This is a study based on data from a prospective longitudinal cohort study of fetal growth performed at 12 US sites (2009-2013). Project gestational age was confirmed by ultrasound between 8 weeks 0 days and 13 weels 6 days, and up to 6 ultrasounds were performed across gestation. Hypertensive disorders were diagnosed based on 2002 American College of Obstetricians and Gynecologists guidelines and grouped hierarchically as severe preeclampsia (including eclampsia or HELLP [hemolysis, elevated liver enzymes, and low platelet count] syndrome), mild preeclampsia, severe gestational hypertension, mild gestational hypertension, or unspecified hypertension. Women without any hypertensive disorder constituted the normotensive group. Growth curves for estimated fetal weight and individual biometric parameters including biparietal diameter, head circumference, abdominal circumference, and femur and humerus length were calculated for each group using linear mixed models with cubic splines. Global and weekly pairwise comparisons were performed between women with a hypertensive disorder compared with normotensive women to analyze differences while adjusting for confounding variables. Delivery gestational age and birthweights were compared among groups. RESULTS: Of 2462 women analyzed, 2296 (93.3%) were normotensive, 63 (2.6%) had mild gestational hypertension, 54 (2.2%) mild preeclampsia, 32 (1.3%) severe preeclampsia, and 17 (0.7%) unspecified hypertension. Compared with normotensive women, those with severe preeclampsia had estimated fetal weights that were reduced between 22 and 38 weeks (all weekly pairwise values of P < .008). Women with severe preeclampsia compared with those without hypertension also had significantly smaller fetal abdominal circumference between 23-31 and 33-37 weeks' gestation (weekly pairwise values of P < .04). Scattered weekly growth differences were noted on other biometric parameters between these 2 groups. The consistent differences in estimated fetal weight and abdominal circumference were not observed between women with other hypertensive disorders and those who were normotensive. Women with severe preeclampsia delivered significantly earlier (mean gestational age 35.9 ± 3.2 weeks) than the other groups (global P < .0001). Birthweights in the severe preeclampsia group were also significantly lower (mean -949.5 g [95% confidence interval, -1117.7 to -781.2 g]; P < .0001) than in the normotensive group. CONCLUSION: Among women with pregnancy-associated hypertensive disorders, only those destined to develop severe preeclampsia demonstrated a significant and consistent difference in fetal growth (ie, smaller estimated fetal weight and abdominal circumference) when compared with normotensive women.

A correlated Bayesian rank likelihood approach to multiple ROC curves for endometriosis.

In analysis of diagnostic data with multiple tests, it is often the case that these tests are correlated. Modeling the correlation explicitly not only produces valid inference results but also enables borrowing of information. Motivated by the Physician Reliability Study (PRS) that investigated the diagnostic performance of physicians in diagnosing endometriosis, we construct a correlated modeling framework to estimate ROC curves and the associated area under the curves. This correlated approach is quite appealing for the PRS data set that suffers from the problem of small sample sizes, as it enables information borrowing between physician groups and sessions. Given that the test scores appear to be non-normal even after logarithm transformation, we use the ranks of the data to conduct likelihood estimation and inference. We use the deviance information criterion to select competing models and conduct simulation studies to assess model performances. In application to the PRS data set, we found that the physicians are not significantly different in their diagnostic performance between groups; however, they are different between the sessions. This suggests that clinical information may play a more important role in physicians' diagnostic performance than their experiences. Our empirical evidence also demonstrates that when using both woman- and physician-specific random effects, the model parameter estimates are much smoother.

Fetal growth velocity: the NICHD fetal growth studies.

BACKGROUND: Accurately identifying pregnancies with accelerated or diminished fetal growth is challenging and generally based on cross-sectional percentile estimates of fetal weight. Longitudinal growth velocity might improve identification of abnormally grown fetuses. OBJECTIVE: We sought to complement fetal size standards with fetal growth velocity, develop a model to compute fetal growth velocity percentiles for any given set of gestational week intervals, and determine association between fetal growth velocity and birthweight. STUDY DESIGN: This was a prospective cohort study with data collected at 12 US sites (2009 through 2013) from 1733 nonobese, low-risk pregnancies included in the singleton standard. Following a standardized sonogram at 10w0d-13w6d, each woman was randomized to 1 of 4 follow-up visit schedules with 5 additional study sonograms (targeted ranges: 16-22, 24-29, 30-33, 34-37, and 38-41 weeks). Study visits could occur ± 1 week from the targeted GA. Ultrasound biometric measurements included biparietal diameter, head circumference, abdominal circumference, and femur length, and estimated fetal weight was calculated. We used linear mixed models with cubic splines for the fixed effects and random effects to flexibly model ultrasound trajectories. We computed velocity percentiles in 2 ways: (1) difference between 2 consecutive weekly measurements (ie, weekly velocity), and (2) difference between any 2 ultrasounds at a clinically reasonable difference between 2 gestational ages (ie, velocity calculator). We compared correlation between fetal growth velocity percentiles and estimated fetal weight percentiles at 4-week intervals, with 32 (±1) weeks' gestation for illustration. Growth velocity was computed as estimated fetal growth rate (g/wk) between ultrasound at that gestational age and from prior visit [ie, for 28-32 weeks' gestational age: velocity = (estimated fetal weight 32-28)/(gestational age 32-28)]. We examined differences in birthweight by whether or not estimated fetal weight and estimated fetal weight velocity were <5th or ≥5th percentiles using χ2. RESULTS: Fetal growth velocity was nonmonotonic, with acceleration early in pregnancy, peaking at 13, 14, 15, and 16 weeks for biparietal diameter, head circumference, femur length, and abdominal circumference, respectively. Biparietal diameter, head circumference, and abdominal circumference had a second acceleration at 19-22, 19-21, and 27-31 weeks, respectively. Estimated fetal weight velocity peaked around 35 weeks. Fetal growth velocity varied slightly by race/ethnicity although comparisons reflected differences for parameters at various gestational ages. Estimated fetal weight velocity percentiles were not highly correlated with fetal size percentiles (Pearson r = 0.40-0.41, P < .001), suggesting that these measurements reflect different aspects of fetal growth and velocity may add additional information to a single measure of estimated fetal weight. At 32 (SD ± 1) weeks, if both estimated fetal weight velocity and size were <5th percentile, mean birthweight was 2550 g; however, even when size remained <5th percentile but velocity was ≥5th percentile, birthweight increased to 2867 g, reflecting the important contribution of higher growth velocities. For estimated fetal weight ≥5th percentile, but growth velocity <5th, birthweight was smaller (3208 vs 3357 g, respectively, P < .001). CONCLUSION: We provide fetal growth velocity data to complement our previous work on fetal growth size standards, and have developed a calculator to compute fetal growth velocity. Preliminary findings suggest that growth velocity adds additional information over knowing fetal size alone.