In this data article, we learn clinical order patterns from inpatient electronic health record (EHR) data at a tertiary academic center from three different cohorts of providers: (1) Clinicians with lower-than-expected patient mortality rates, (2) clinicians with higher-than-expected patient mortality rates, and (3) an unfiltered population of clinicians. We extract and make public these order patterns learned from each clinician cohort associated with six common admission diagnoses (e.g. pneumonia, chest pain, etc.). We also share a reusable reference standard or benchmark for evaluating automatically-learned clinical order patterns for each admission diagnosis, based on a manual review of clinical practice literature. The data shared in this article can support further study, evaluation, and translation of data-driven CDS systems. Further interpretation and discussion of this data can be found in Wang et al. (2018).

Uncertainty and variability is pervasive in medical decision making with insufficient evidence-based medicine and inconsistent implementation where established knowledge exists. Clinical decision support constructs like order sets help distribute expertise, but are constrained by knowledge-based development. We previously produced a data-driven order recommender system to automatically generate clinical decision support content from structured electronic medical record data on >19K hospital patients. We now present the first structured validation of such automatically generated content against an objective external standard by assessing how well the generated recommendations correspond to orders referenced as appropriate in clinical practice guidelines. For example scenarios of chest pain, gastrointestinal hemorrhage, and pneumonia in hospital patients, the automated method identifies guideline reference orders with ROC AUCs (c-statistics) (0.89, 0.95, 0.83) that improve upon statistical prevalence benchmarks (0.76, 0.74, 0.73) and pre-existing human-expert authored order sets (0.81, 0.77, 0.73) (P<10(-30) in all cases). We demonstrate that data-driven, automatically generated clinical decision support content can reproduce and optimize top-down constructs like order sets while largely avoiding inappropriate and irrelevant recommendations. This will be even more important when extrapolating to more typical clinical scenarios where well-defined external standards and decision support do not exist.

Treatment of cancer has been revolutionized by immune checkpoint blockade therapies. Despite the high rate of response in advanced melanoma, the majority of patients succumb to disease. To identify factors associated with success or failure of checkpoint therapy, we profiled transcriptomes of 16,291 individual immune cells from 48 tumor samples of melanoma patients treated with checkpoint inhibitors. Two distinct states of CD8+ T cells were defined by clustering and associated with patient tumor regression or progression. A single transcription factor, TCF7, was visualized within CD8+ T cells in fixed tumor samples and predicted positive clinical outcome in an independent cohort of checkpoint-treated patients. We delineated the epigenetic landscape and clonality of these T cell states and demonstrated enhanced antitumor immunity by targeting novel combinations of factors in exhausted cells. Our study of immune cell transcriptomes from tumors demonstrates a strategy for identifying predictors, mechanisms, and targets for enhancing checkpoint immunotherapy.

Cell fate decisions occur through the switch-like, irreversible activation of fate-specifying genes. These activation events are often assumed to be tightly coupled to changes in upstream transcription factors, but could also be constrained by cis-epigenetic mechanisms at individual gene loci. Here, we studied the activation of Bcl11b, which controls T-cell fate commitment. To disentangle cis and trans effects, we generated mice where two Bcl11b copies are tagged with distinguishable fluorescent proteins. Quantitative live microscopy of progenitors from these mice revealed that Bcl11b turned on after a stochastic delay averaging multiple days, which varied not only between cells but also between Bcl11b alleles within the same cell. Genetic perturbations, together with mathematical modeling, showed that a distal enhancer controls the rate of epigenetic activation, while a parallel Notch-dependent trans-acting step stimulates expression from activated loci. These results show that developmental fate transitions can be controlled by stochastic cis-acting events on individual loci.

The relationship of SARS-CoV-2 lung infection and severity of pulmonary disease is not fully understood. We analyzed autopsy specimens from 24 patients who succumbed to SARS-CoV-2 infection using a combination of different RNA and protein analytical platforms to characterize inter- and intra- patient heterogeneity of pulmonary virus infection. There was a spectrum of high and low virus cases that was associated with duration of disease and activation of interferon pathway genes. Using a digital spatial profiling platform, the virus corresponded to distinct spatial expression of interferon response genes and immune checkpoint genes demonstrating the intra-pulmonary heterogeneity of SARS-CoV-2 infection.

Immune checkpoint inhibitors (ICIs) show promise, but most patients do not respond. We identify and validate biomarkers from extracellular vesicles (EVs), allowing non-invasive monitoring of tumor- intrinsic and host immune status, as well as a prediction of ICI response. We undertook transcriptomic profiling of plasma-derived EVs and tumors from 50 patients with metastatic melanoma receiving ICI, and validated with an independent EV-only cohort of 30 patients. Plasma-derived EV and tumor transcriptomes correlate. EV profiles reveal drivers of ICI resistance and melanoma progression, exhibit differentially expressed genes/pathways, and correlate with clinical response to ICI. We created a Bayesian probabilistic deconvolution model to estimate contributions from tumor and non-tumor sources, enabling interpretation of differentially expressed genes/pathways. EV RNA-seq mutations also segregated ICI response. EVs serve as a non-invasive biomarker to jointly probe tumor-intrinsic and immune changes to ICI, function as predictive markers of ICI responsiveness, and monitor tumor persistence and immune activation.

This study assesses whether neural networks trained on electronic health record (EHR) data can anticipate what individual clinical orders and existing institutional order set templates clinicians will use more accurately than existing decision support tools.

The impact of existing malnutrition on stroke outcomes is poorly recognised and treated. Evidence was systematically reviewed and quantified by meta-analysis.

More than 10% of the world's population is chronically infected with HIV, hepatitis C virus (HCV) or hepatitis B virus (HBV), all of which can cause severe disease and death. These viruses persist in part because continuous antigenic stimulation causes the deterioration of virus-specific cytotoxic T lymphocyte (CTL) function and survival. Additionally, antiviral CTLs autonomously suppress their responses to limit immunopathology by upregulating inhibitory receptors such as programmed cell death 1 (PD-1). Identification and blockade of the pathways that induce CTL dysfunction may facilitate the clearance of chronic viral infections. We found that the prostaglandin E2 (PGE₂) receptors EP2 and EP4 were upregulated on virus-specific CTLs during chronic lymphocytic choriomeningitis virus (LCMV) infection and suppressed CTL survival and function. We show that the combined blockade of PGE₂ and PD-1 signaling was therapeutic in terms of improving viral control and augmenting the numbers of functional virus-specific CTLs. Thus, PGE₂ inhibition is both an independent candidate therapeutic target and a promising adjunct therapy to PD-1 blockade for the treatment of HIV and other chronic viral infections.

Protein kinase B (also known as AKT) and the mechanistic target of rapamycin (mTOR) are central regulators of T cell differentiation, proliferation, metabolism, and survival. Here, we show that during chronic murine lymphocytic choriomeningitis virus infection, activation of AKT and mTOR are impaired in antiviral cytotoxic T lymphocytes (CTLs), resulting in enhanced activity of the transcription factor FoxO1. Blockade of inhibitory receptor programmed cell death protein 1 (PD-1) in vivo increased mTOR activity in virus-specific CTLs, and its therapeutic effects were abrogated by the mTOR inhibitor rapamycin. FoxO1 functioned as a transcriptional activator of PD-1 that promoted the differentiation of terminally exhausted CTLs. Importantly, FoxO1-null CTLs failed to persist and control chronic viral infection. Collectively, this study shows that CTLs adapt to persistent infection through a positive feedback pathway (PD-1?FoxO1?PD-1) that functions to both desensitize virus-specific CTLs to antigen and support their survival during chronic viral infection.

Lineage tracing provides key insights into the fate of individual cells in complex organisms. Although effective genetic labeling approaches are available in model systems, in humans, most approaches require detection of nuclear somatic mutations, which have high error rates, limited scale, and do not capture cell state information. Here, we show that somatic mutations in mtDNA can be tracked by single-cell RNA or assay for transposase accessible chromatin (ATAC) sequencing. We leverage somatic mtDNA mutations as natural genetic barcodes and demonstrate their utility as highly accurate clonal markers to infer cellular relationships. We track native human cells both in vitro and in vivo and relate clonal dynamics to gene expression and chromatin accessibility. Our approach should allow clonal tracking at a 1,000-fold greater scale than with nuclear genome sequencing, with simultaneous information on cell state, opening the way to chart cellular dynamics in human health and disease.

Overcoming intrinsic resistance to immune checkpoint blockade for microsatellite stable (MSS) colorectal cancer (CRC) and pancreatic ductal adenocarcinoma (PDAC) remains challenging. We conducted a single-arm, non-randomized, phase II trial (NCT03104439) combining radiation, ipilimumab and nivolumab to treat patients with metastatic MSS CRC (n = 40) and PDAC (n = 25) with an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1. The primary endpoint was disease control rate (DCR) by intention to treat. DCRs were 25% for CRC (ten of 40; 95% confidence interval (CI), 13-41%) and 20% for PDAC (five of 25; 95% CI, 7-41%). In the per-protocol analysis, defined as receipt of radiation, DCR was 37% (ten of 27; 95% CI, 19-58%) in CRC and 29% (five of 17; 95% CI, 10-56%) in PDAC. Pretreatment biopsies revealed low tumor mutational burden for all samples but higher numbers of natural killer (NK) cells and expression of the HERVK repeat RNA in patients with disease control. This study provides proof of concept of combining radiation with immune checkpoint blockade in immunotherapy-resistant cancers.

While BRAF inhibitor combinations with EGFR and/or MEK inhibitors have improved clinical efficacy in BRAFV600E colorectal cancer (CRC), response rates remain low and lack durability. Preclinical data suggest that BRAF/MAPK pathway inhibition may augment the tumor immune response. We performed a proof-of-concept single-arm phase 2 clinical trial of combined PD-1, BRAF and MEK inhibition with sparatlizumab (PDR001), dabrafenib and trametinib in 37 patients with BRAFV600E CRC. The primary end point was overall response rate, and the secondary end points were progression-free survival, disease control rate, duration of response and overall survival. The study met its primary end point with a confirmed response rate (24.3% in all patients; 25% in microsatellite stable patients) and durability that were favorable relative to historical controls of BRAF-targeted combinations alone. Single-cell RNA sequencing of 23 paired pretreatment and day 15 on-treatment tumor biopsies revealed greater induction of tumor cell-intrinsic immune programs and more complete MAPK inhibition in patients with better clinical outcome. Immune program induction in matched patient-derived organoids correlated with the degree of MAPK inhibition. These data suggest a potential tumor cell-intrinsic mechanism of cooperativity between MAPK inhibition and immune response, warranting further clinical evaluation of optimized targeted and immune combinations in CRC. ClinicalTrials.gov registration: NCT03668431.

To develop prediction models for intensive care unit (ICU) vs non-ICU level-of-care need within 24 hours of inpatient admission for emergency department (ED) patients using electronic health record data.

Metastatic pancreatic adenocarcinoma remains one of the deadliest cancer diagnoses with 5-year survival rates as low as 3%. For decades, gemcitabine remained the mainstay of systemic therapy before the approvals of FOLFIRINOX and gemcitabine with nab-paclitaxel. Despite these advances in the early 2010s, almost all patients progress on systemic chemotherapy and significant effort is needed to identify novel therapeutic targets. A promising array of approaches is currently under investigation, enabled by deeper understanding of the immune system within the tumor microenvironment (TME) and of the key vulnerabilities in pathways essential for tumor survival. In this review, we will explore the different approaches to boost tumor immunity and to target tumor metabolic pathways that are currently under clinical investigation for systemic treatment, and highlight the promising therapeutic areas that may give rise to the next generation of therapies for pancreatic cancer.

Regulatory relationships between transcription factors (TFs) and their target genes lie at the heart of cellular identity and function; however, uncovering these relationships is often labor-intensive and requires perturbations. Here, we propose a principled framework to systematically infer gene regulation for all TFs simultaneously in cells at steady state by leveraging the intrinsic variation in the transcriptional abundance across single cells. Through modeling and simulations, we characterize how transcriptional bursts of a TF gene are propagated to its target genes, including the expected ranges of time delay and magnitude of maximum covariation. We distinguish these temporal trends from the time-invariant covariation arising from cell states, and we delineate the experimental and technical requirements for leveraging these small but meaningful cofluctuations in the presence of measurement noise. While current technology does not yet allow adequate power for definitively detecting regulatory relationships for all TFs simultaneously in cells at steady state, we investigate a small-scale dataset to inform future experimental design. This study supports the potential value of mapping regulatory connections through stochastic variation, and it motivates further technological development to achieve its full potential.

Though artificial intelligence (AI) has the potential to augment the patient-physician relationship in primary care, bias in intelligent health care systems has the potential to differentially impact vulnerable patient populations.

MicroRNAs have emerged as key regulators of B cell fate decisions and immune function. Deregulation of several microRNAs in B cells leads to the development of autoimmune disease and cancer in mice. We demonstrate that the microRNA-212/132 cluster (miR-212/132) is induced in B cells in response to B cell receptor signaling. Enforced expression of miR-132 results in a block in early B cell development at the prepro-B cell to pro-B cell transition and induces apoptosis in primary bone marrow B cells. Importantly, loss of miR-212/132 results in accelerated B cell recovery after antibody-mediated B cell depletion. We find that Sox4 is a target of miR-132 in B cells. Co-expression of SOX4 with miR-132 rescues the defect in B cell development from overexpression of miR-132 alone, thus suggesting that miR-132 may regulate B lymphopoiesis through Sox4. In addition, we show that the expression of miR-132 can inhibit cancer development in cells that are prone to B cell cancers, such as B cells expressing the c-Myc oncogene. We have thus uncovered miR-132 as a novel contributor to B cell development.

Chemotherapy is ubiquitous in first-line treatment of advanced gastric cancer, yet responses are heterogeneous, and little is known about mediators of chemotherapy response. To move forward, an understanding of the effects of standard chemotherapy on the tumor-immune microenvironment (TME) is needed. Coupling whole-exome sequencing, bulk RNA and single-cell transcriptomics from paired pretreatment and on-treatment samples in treatment-naïve patients with HER2-positive and HER2-negative gastric cancer, we define features associated with response to platinum-based chemotherapy. Response was associated with on-treatment TME remodeling including natural killer (NK) cell recruitment, decreased tumor-associated macrophages, M1-macrophage repolarization, and increased effector T-cell infiltration. Among chemotherapy nonresponders, we observed low/absent PD-L1 expression or modulation, on-treatment increases in Wnt signaling, B-cell infiltration, and LAG3-expressing T cells coupled to an exodus of dendritic cells. We did not observe significant genomic changes in early on-treatment sampling. We provide a map of on-treatment TME modulation with standard chemotherapy and nominate candidate future approaches.

The study sought to determine whether machine learning can predict initial inpatient total daily dose (TDD) of insulin from electronic health records more accurately than existing guideline-based dosing recommendations.