Conditioning in 2 x 2 tables

Enrichment of melanoma-associated T cells in 6-thioguanine-resistant T cells from metastatic melanoma patients

This study examines whether 6-thioguanine resistant T cells (mutant) from metastatic melanoma patients are enriched for melanoma-associated T cells compared to T cells obtained analogously without thioguanine selection (wild-type). Melanoma-associated antigen pentamer staining was performed on 5 tumour and 9 peripheral blood samples from metastatic melanoma patients. T cell receptor beta chain repertoire was examined via Sanger sequencing of mutant and wild-type in blood and tumour from metastatic melanoma patients at times of tumour progression (n = 8) and via Illumina sequencing in tumour derived T cells and in uncultured T cells (uncultured), wild-type and mutant from blood before and after immune checkpoint blockade (n = 1). Mutant from tumour (3 of 5; P < 0.001), but not blood (0 of 9), were enriched compared to wild-type for binding melanoma-associated antigen pentamers. T cell receptor beta analysis in patients with tumour progression (n = 8) detected increased melanoma associated T cells in mutant compared to wild-type from blood (Monte Carlo P = 10). Comparison of blood samples before and after immune checkpoint blockade with prior tumor from one metastatic melanoma patient detected increased T cell receptor beta sharing between tumour and mutant compared to tumour and wild-type or tumour and uncultured: 11.0% (72/656), 1.5% (206/13 639) and 1.3% (381/29 807), respectively (Monte Carlo P = 10 for mutant versus wild-type and mutant versus uncultured). These data demonstrate that mutant in metastatic melanoma patients are enriched for melanoma-associated T cells and are candidate probes to study in vivo melanoma-reactive T cells.

Re-randomization tests for unplanned changes in clinical trials

Despite the best efforts of investigators, problems forcing design changes can occur in clinical trials. Changes are usually relatively minor, but sometimes not. The primary endpoint or analysis may need to be revised, for example. It is common to regard any conclusion from such a tarnished trial as hypothesis-generating rather than definitive. This article reviews a very useful technique, re-randomization tests, for dealing with such anomalies. Re-randomization tests remain valid for testing a strong null hypothesis that treatment has no effect on the data that led to design changes. Another way of expressing this is that the data used to inform a design change must give no information about the treatment labels. This restriction has implications for limiting the amount of information examined by a committee deciding whether to make design alterations. While nothing can eliminate the pall cast by breaches of protocol, re-randomization tests following blinded and limited data examination go a long way toward amelioration.

Adverse Cardiovascular Response to Aerobic Exercise Training: Is This a Concern?

PURPOSE

Aerobic exercise training in sedentary individuals improves physical fitness and various cardiovascular (CV) biomarkers. Nevertheless, there has been controversy as to whether exercise training may adversely affect some biomarkers in a small segment of the population. The purpose of this study was to investigate whether clinically significant worsening of CV biomarkers was more prevalent among individuals randomized to a supervised endurance training program as compared with those randomized to a control condition.

METHODS

Baseline and end of study measurements of fasting insulin (FI), triglycerides (TG), resting systolic blood pressure (SBP), and HDL cholesterol (HDL-C) were obtained on 1188 healthy sedentary subjects from 4 clinical studies. Each study randomized subjects to 4- to 6-month supervised aerobic exercise programs or to a control group of no supervised exercise training. For each of the 4 CV biomarkers, we calculated the respective proportions of control and exercise group subjects whose baseline-to-follow-up changes were greater than or equal to previously reported adverse change (AC) thresholds. Those thresholds were increases of 24 pmol · L(-1) or greater for FI, 0.42 mmol · L(-1) or greater for TG, 10 mm Hg or greater for SBP, and a decrease of 0.12 mmol · L(-1) or greater for HDL-C.

RESULTS

The respective proportions of subjects meeting the AC threshold in the control and exercise groups were 15.2% versus 9.6% (P = 0.02) for FI, 14.9% versus 13.1% (P = 0.37) for TG, 16.9% versus 15.8% (P = 0.52) for SBP, and 28.6% versus 22.5% (P = 0.03) for HDL-C. All were nonsignificant at the 0.0125 Bonferroni threshold adjusting for multiple comparisons.

CONCLUSIONS

These findings do not support the concept that aerobic exercise training increases the risk of adverse changes in the CV biomarkers we studied.

COMPASS identifies T-cell subsets correlated with clinical outcomes

Advances in flow cytometry and other single-cell technologies have enabled high-dimensional, high-throughput measurements of individual cells as well as the interrogation of cell population heterogeneity. However, in many instances, computational tools to analyze the wealth of data generated by these technologies are lacking. Here, we present a computational framework for unbiased combinatorial polyfunctionality analysis of antigen-specific T-cell subsets (COMPASS). COMPASS uses a Bayesian hierarchical framework to model all observed cell subsets and select those most likely to have antigen-specific responses. Cell-subset responses are quantified by posterior probabilities, and human subject-level responses are quantified by two summary statistics that describe the quality of an individual's polyfunctional response and can be correlated directly with clinical outcome. Using three clinical data sets of cytokine production, we demonstrate how COMPASS improves characterization of antigen-specific T cells and reveals cellular 'correlates of protection/immunity' in the RV144 HIV vaccine efficacy trial that are missed by other methods. COMPASS is available as open-source software.

Mixture models for single-cell assays with applications to vaccine studies

Blood and tissue are composed of many functionally distinct cell subsets. In immunological studies, these can be measured accurately only using single-cell assays. The characterization of these small cell subsets is crucial to decipher system-level biological changes. For this reason, an increasing number of studies rely on assays that provide single-cell measurements of multiple genes and proteins from bulk cell samples. A common problem in the analysis of such data is to identify biomarkers (or combinations of biomarkers) that are differentially expressed between two biological conditions (e.g. before/after stimulation), where expression is defined as the proportion of cells expressing that biomarker (or biomarker combination) in the cell subset(s) of interest. Here, we present a Bayesian hierarchical framework based on a beta-binomial mixture model for testing for differential biomarker expression using single-cell assays. Our model allows the inference to be subject specific, as is typically required when assessing vaccine responses, while borrowing strength across subjects through common prior distributions. We propose two approaches for parameter estimation: an empirical-Bayes approach using an Expectation-Maximization algorithm and a fully Bayesian one based on a Markov chain Monte Carlo algorithm. We compare our method against classical approaches for single-cell assays including Fisher's exact test, a likelihood ratio test, and basic log-fold changes. Using several experimental assays measuring proteins or genes at single-cell level and simulations, we show that our method has higher sensitivity and specificity than alternative methods. Additional simulations show that our framework is also robust to model misspecification. Finally, we demonstrate how our approach can be extended to testing multivariate differential expression across multiple biomarker combinations using a Dirichlet-multinomial model and illustrate this approach using single-cell gene expression data and simulations.