Prioritizing HIV comparative effectiveness trials based on value of information: generic versus brand-name ART in the US

Value of Information (VOI) analysis examines whether to acquire information before making a decision. We introduced VOI to the HIV audience, using the example of generic antiretroviral therapy (ART) in the US.

Methods and Findings:We used a mathematical model and probabilistic sensitivity analysis (PSA) to generate probability distributions of survival (in quality-adjusted life years, QALYs) and cost for three potential first-line ART regimens: three-pill generic, two-pill generic, and single-pill branded. These served as input for a comparison of two hypothetical two-arm trials: three-pill generic versus single-pill branded; and two-pill generic versus single-pill branded. We modeled pre-trial uncertainty by defining probability distributions around key inputs, including 24-week HIV-RNA suppression and subsequent ART failure. We assumed that, without a trial, patients received the single-pill branded strategy. Post-trial, we assumed that patients received the most cost-effective strategy. For both trials, we quantified the probability of changing to a generic-based regimen upon trial completion and the expected VOI in terms of improved health outcomes and costs. Assuming a willingness to pay (WTP) threshold of $100?000/QALY, the three-pill trial led to more treatment changes (84%) than the two-pill trial (78%). Estimated VOI was $48?000 (three-pill trial) and $35?700 (two-pill trial) per future patient initiating ART.

Conclusions:A three-pill trial of generic ART is more likely to lead to post-trial treatment changes and to provide more value than a two-pill trial if policy decisions are based on cost-effectiveness. Value of Information analysis can identify trials likely to confer the greatest impact and value for HIV care.     

Building bridges across electronic health record systems through inferred phenotypic topics

ObjectiveData in electronic health records (EHRs) is being increasingly leveraged for secondary uses, ranging from biomedical association studies to comparative effectiveness. To perform studies at scale and transfer knowledge from one institution to another in a meaningful way, we need to harmonize the phenotypes in such systems. Traditionally, this has been accomplished through expert specification of phenotypes via standardized terminologies, such as billing codes. However, this approach may be biased by the experience and expectations of the experts, as well as the vocabulary used to describe such patients. The goal of this work is to develop a data-driven strategy to (1) infer phenotypic topics within patient populations and (2) assess the degree to which such topics facilitate a mapping across populations in disparate healthcare systems.MethodsWe adapt a generative topic modeling strategy, based on latent Dirichlet allocation, to infer phenotypic topics. We utilize a variance analysis to assess the projection of a patient population from one healthcare system onto the topics learned from another system. The consistency of learned phenotypic topics was evaluated using (1) the similarity of topics, (2) the stability of a patient population across topics, and (3) the transferability of a topic across sites. We evaluated our approaches using four months of inpatient data from two geographically distinct healthcare systems: (1) Northwestern Memorial Hospital (NMH) and (2) Vanderbilt University Medical Center (VUMC).ResultsThe method learned 25 phenotypic topics from each healthcare system. The average cosine similarity between matched topics across the two sites was 0.39, a remarkably high value given the very high dimensionality of the feature space. The average stability of VUMC and NMH patients across the topics of two sites was 0.988 and 0.812, respectively, as measured by the Pearson correlation coefficient. Also the VUMC and NMH topics have smaller variance of characterizing patient population of two sites than standard clinical terminologies (e.g., ICD9), suggesting they may be more reliably transferred across hospital systems.ConclusionsPhenotypic topics learned from EHR data can be more stable and transferable than billing codes for characterizing the general status of a patient population. This suggests that EHR-based research may be able to leverage such phenotypic topics as variables when pooling patient populations in predictive models.     

Combining two strategies to improve perfusion and drug delivery in solid tumors

Blood perfusion in tumors can be significantly lower than that in the surrounding normal tissue owing to the leakiness and/or compression of tumor blood vessels. Impaired perfusion reduces oxygen supply and results in a hypoxic microenvironment. Hypoxia promotes tumor progression and immunosuppression, and enhances the invasive and metastatic potential of cancer cells. Furthermore, poor perfusion lowers the delivery of systemically administered drugs. Therapeutic strategies to improve perfusion include reduction in vascular permeability by vascular normalization and vascular decompression by alleviating physical forces (solid stress) inside tumors. Both strategies have shown promise, but guidelines on how to use these strategies optimally are lacking. To this end, we developed a mathematical model to guide the optimal use of these strategies. The model accounts for vascular, transvascular, and interstitial fluid and drug transport as well as the diameter and permeability of tumor vessels. Model simulations reveal an optimal perfusion region when vessels are uncompressed, but not very leaky. Within this region, intratumoral distribution of drugs is optimized, particularly for drugs 10 nm in diameter or smaller and of low binding affinity. Therefore, treatments should modify vessel diameter and/or permeability such that perfusion is optimal. Vascular normalization is more effective for hyperpermeable but largely uncompressed vessels (e.g., glioblastomas), whereas solid stress alleviation is more beneficial for compressed but less-permeable vessels (e.g., pancreatic ductal adenocarcinomas). In the case of tumors with hyperpermeable and compressed vessels (e.g., subset of mammary carcinomas), the two strategies need to be combined for improved treatment outcomes.Perfused vessels are necessary for enabling adequate oxygenation and distribution of systemically administered drugs in solid tumors. However, perfusion rates in some regions of a tumor can be significantly lower than that in the peritumor normal tissue, leading to hypoxia, low pH, and inadequate drug delivery. Impaired blood perfusion in tumors could result from (i) excessive fluid loss from the vasculature to the interstitial space owing to vessel hyperpermeability (Fig. 1A), (ii) increased resistance to fluid flow caused by vessel tortuosity, and (iii) reduced effective cross-sectional area for blood flow due to vessel compression (Fig. 1C) (1, 2). Vessel hyperpermeability and tortuosity can be lowered using judicious doses of antiangiogenic agents (1, 3–6). Fig. 1 A and B shows a schematic of how vascular normalization can improve perfusion by improving the structure and composition of the vessel wall, and Open in a separate windowFig. 1.Schematic of therapeutic strategies to improve tumor perfusion. (A) Abnormalities in interendothelial junctions, pericyte coverage, and/or basement membrane lead to hyperpermeability of tumor blood vessels and excessive fluid leakage that slows down blood flow. (B) Vascular normalization fortifies the vessel wall, resulting in smaller interendothelial gaps (“pores”) and improved perfusion. (C) Structural components of the tumor microenvironment exert forces on blood vessels, resulting in vessel compression and reduced blood flow. (D) Alleviation of these forces by selective depletion of tumor constituents (e.g., cells or extracellular matrix) can decompress the vessels and improve vessel perfusion. BM, basement membrane; CC, cancer and/or stromal cell; EC, endothelial cell; ECM, extracellular matrix; PC, pericyte.

Table 1.

Studies showing vascular normalization improves perfusion measured as improved oxygenation

Nasogastric Aspiration as an Indicator for Feed Absorption in Model-Based Glycemic Control in Neonatal Intensive Care

Background

STAR (stochastic targeted) is a glycemic control model-based framework for critically ill neonates that has shown benefits in reducing hypoglycemia and hyperglycemia. STAR uses a stochastic matrix method to forecast future changes in a patient’s insulin sensitivity and then applies this result to a physiological model to select an optimal insulin treatment. Nasogastric aspiration may be used as an indicator to suggest periods of care when enteral feed absorption is compromised, improving the performance of glycemic control. An analysis has been carried out to investigate the effect of poorly absorbed feeds on glycemic control.

Method

Clinical data were collected from eight patients on insulin therapy and enteral feed, which included large or significantly milky aspirates. Patients had a median gestational age of 25 weeks and postnatal age of 5.5 days. Virtual patients were created using the NICING model, and insulin sensitivity (S_I) profiles were fit. Alternative feed profiles were generated whereby enteral feed absorption was redistributed with time to account for poor feed absorption. The effect of poor feed absorption, as indicated by aspirates, is investigated.

Results

The average percentage change of S_I 4 h before a significant aspirate was 1.16%, and 1.49% in the 4 h following the aspirate. No distinct relationship was found between the fractional change in S_I and the volume of the aspirate. Accounting for aspirates had a clinically negligible impact on glycemic control in virtual trials.

Conclusion

Accounting for aspirates by manipulating enteral feed profiles had a minimal influence on both modeling and controlling glycemia in neonates. The impact of this method is clinically insignificant, suggesting that a population constant for the rate of glucose absorption in the gut adequately models feed absorption within the STAR framework.J Diabetes Sci Technol 2013;7(3):717–726     

Determination of virus burst size in vivo using a single-cycle SIV in rhesus macaques

A single-cycle simian immunodeficiency virus (scSIV) that undergoes only one round of infection and replication was constructed to calculate the total number of virons produced by an SIV-infected cell in vivo. Four Mamu-A*01 rhesus macaques were inoculated on two occasions 11 weeks apart with the scSIV by ex vivo infection and i.v. reinfusion of autologous cells. After each inoculation, plasma viral loads peaked between 1 and 2.5 days and then declined exponentially in one or two phases to below detection limits within 2 weeks. Although higher levels of SIV-specific cytotoxic T lymphocytes and modest increases in antibody responses were observed for each animal after the second inoculation, decay rates of the infected cells were only minimally affected. Analyzing the viral load data with a mathematical model, the in vivo viral burst size averaged 4.0 x 10(4) and 5.5 x 10(4) virions per cell for the first and second inoculations, respectively, with no significant difference between the two inoculations. This estimate, in conjunction with our prior understanding of other quantitative viral and cellular parameters during SIV and HIV infection, provides critical insights into the dynamic process of viral production and its interplay with the infected host in vivo.     

T84-intestinal epithelial exosomes bear MHC class II/peptide complexes potentiating antigen presentation by dendritic cells

BACKGROUND & AIMS: Intestinal epithelial cells release antigen-presenting vesicles (exosomes) bearing major histocompatibility complex class II/peptide complexes stimulating specific immune responses in vivo. To characterize further the role of human epithelial exosomes in antigen presentation, their capacity to load antigenic peptides, bind immune target cells, and induce T-cell activation was analyzed in vitro. METHODS: The capacity of exosomes derived from the HLA-DR4-expressing, intestinal epithelial cell line T84 to load the HLA-DR4-specific peptide (3)H-HSA 64-76 and to activate a HLA-DR4-restricted T-cell hybridoma was tested in the presence or absence of human monocyte-derived dendritic cells (DCs). Interaction of fluorescein isothiocyanate-labeled exosomes with T cells and DCs was analyzed by flow cytometry and confocal microscopy. RESULTS: T84-derived exosomes, enriched in CD9, CD81, CD82, and A33 antigen, were capable of binding specifically human serum albumin (HSA) 64-76 peptide on HLA-DR4 molecules and of interacting preferentially with DCs. HSA-loaded exosomes were unable to activate the T-cell hybridoma directly but induced a productive T-cell activation through DCs. When HSA peptide was bound to exosomal HLA-DR4 molecules instead of in a soluble form, the threshold of peptide presentation by DCs was markedly decreased (x10(-3)). CONCLUSIONS: Exosomes released by intestinal epithelial cells bear exogenous peptides complexed to major histocompatibility complex class II molecules and interact preferentially with DCs, strongly potentiating peptide presentation to T cells. Epithelial exosomes constitute a powerful link between luminal antigens and local immune cells by mediating the transfer of tiny amounts of luminal antigenic information and facilitating immune surveillance at mucosal surfaces.     

Defining functional distance using manifold embeddings of gene ontology annotations

Although rigorous measures of similarity for sequence and structure are now well established, the problem of defining functional relationships has been particularly daunting. Here, we present several manifold embedding techniques to compute distances between Gene Ontology (GO) functional annotations and consequently estimate functional distances between protein domains. To evaluate accuracy, we correlate the functional distance to the well established measures of sequence, structural, and phylogenetic similarities. Finally, we show that manual classification of structures into folds and superfamilies is mirrored by proximity in the newly defined function space. We show how functional distances place structure-function relationships in biological context resulting in insight into divergent and convergent evolution. The methods and results in this paper can be readily generalized and applied to a wide array of biologically relevant investigations, such as accuracy of annotation transference, the relationship between sequence, structure, and function, or coherence of expression modules.     

Integration from proteins to organs: the IUPS Physiome Project

The IUPS Physiome Project is an internationally collaborative open source project intended to provide a public domain framework for computational physiology, including the development of modeling standards, computational tools and web-accessible databases of models of structure and function at all spatial scales and across all organ systems. Here, we illustrate the application of this multi-scale modeling approach to three organ systems: the heart, the lungs and the musculo-skeletal system, and in each case we show how the organ level models incorporate tissue and cell-level physiology. Although the computational physiology framework presented here does not yet incorporate models of ageing processes, the model-based approach is certainly capable of describing ageing and disease-related processes both via parameter changes within the models of normal physiological processes and via models of additional processes added to the framework.     

Comparative analysis of the fate of donor dendritic cells and B cells and their influence on alloreactive T cell responses under tacrolimus immunosuppression

We have shown that tacrolimus (TAC)-induced liver allograft acceptance is associated with migration and persistence of donor B cells and dendritic cells (DC). To clarify whether these MHC class II+ leukocytes have favorable roles in inducing tolerance, we analyzed recipient T cell reactions after allogeneic B or DC infusion. LEW rat B cells localized exclusively in BN host B cell follicles without any direct contact with host T cells. While few donor DC migrated to T cell areas and marginal zones, they were captured by host APC, suggesting that allogeneic MHC class II+ cells may induce immune reactions via the indirect pathway. Although DC-infused non-immunosuppressed recipients showed enhanced ex vivo anti-donor responses, persistent in vitro donor-specific hyporeactivity was seen equally with donor DC or B cell infusion under TAC. The results indicate that donor MHC class II+ APC are capable of regulating recipient immune reactions under TAC. Possible involvement of the indirect pathway of allorecognition is discussed.