Tempus Labs (United States)

In the Philippines, approximately 13 million adults smoke. Policies to raise cigarette taxes, increase health warning label (HWL) size, and restrict menthol could reduce smoking. We conducted an online discrete choice experiment (DCE) in November 2023 to assess hypothetical policy responses among 886 Filipino adults who smoke across three regions of the Philippines (Luzon, Visayas, Mindanao). The DCE included ten choice sets of three cigarette packs. Packs varied on excise tax [60 (ref.), 70, 80, 90 pesos (PHP)], packaging [domestic packs: 50% HWL branded (ref.), 50% HWL plain, 85% HWL branded, 85% HWL plain; foreign pack, no HWL], and flavor [tobacco (ref.), menthol]. Participants were randomly assigned to see one choice set using a programmed 1:1 randomization ratio and asked to select which pack: 1) they would choose, 2) would make them most consider quitting, and 3) would be less harmful. Participants had lower odds of choosing packs with a 70 (adjusted odds ratio, AOR=0.58; 95% CI: 0.45-0.74), 80 (AOR=0.33; 95% CI: 0.25-0.43), or 90 (AOR=0.24; 95% CI: 0.18-0.33) PHP excise tax, and higher odds (AOR=4.02; 95% CI: 2.86-5.66) of choosing the foreign branded pack with no HWL. For quitting, participants had higher odds of selecting packs with an 80 (AOR=1.53; 95% CI: 1.18-1.99) or 90 (AOR=1.94; 95% CI: 1.50-2.51) PHP tax and 3.40 (95% CI: 2.58-4.46) and 3.30 (95% CI: 2.42-4.49) higher odds of selecting a branded or plain pack with 85% HWL coverage, respectively. Participants had lower odds of thinking packs with a 70 (AOR=0.69; 95% CI: 0.52-0.92), 80 (AOR=0.73; 95% CI: 0.56-0.96), or 90 (AOR=0.67; 95% CI: 0.50-0.89) PHP excise tax were less harmful and higher odds (AOR=5.21; 95% CI: 3.70-7.32) of thinking the foreign pack was less harmful. Results suggest raising the excise tax to at least 80 PHP and implementing 85% HWLs may encourage smoking cessation in the Philippines. Our findings can be used by policymakers and advocates in the Philippines to implement strong tobacco control policies to protect public health.

ABSTRACTHistorically, clonal hematopoiesis (CH) has been recognized as a confounder of cell‐free DNA (cfDNA) testing. Recent evidence now demonstrates the role of CH as a risk factor in health, generating distinct sources of cfDNA that can be leveraged for liquid biopsy diagnostics. Nonetheless, gaps in standardization challenge the advancement of such diagnostics from development to regulatory approval, through clinical trials, and ultimately, to routine implementation. In 2024, the Blood Profiling Atlas in Cancer (BLOODPAC) Consortium, a collaborative infrastructure for developing standards and best practices for liquid biopsy assays, established the CH/clonal hematopoiesis of indeterminate potential (CHIP) Working Group to address the need for accurate identification and removal of CH from liquid biopsy results. As a first step to support the interpretability of CH/CHIP results, the Working Group developed this lexicon to standardize terms and provide a unified vocabulary related to CH and liquid biopsy, DNA sequencing tests, biomarkers, and clinical use cases, facilitating communication within the field. BLOODPAC's CH/CHIP Working Group believes that terminology agreement across these various stakeholders can improve communication in the field and unify future data collection efforts across studies.

ABSTRACTBackgroundDespite increased awareness of its clinical consequences, personalized risk stratification in patients with moderate aortic stenosis (AS) remains ambiguous. We studied predictors of progression and clinical outcomes in moderate AS to improve risk stratification and add to the existing literature.MethodsData collected through the Tempus Next care pathway intelligence platform at our tertiary referral center were retrospectively analyzed. The reports of echocardiograms performed from October 2017 to January 2020 were screened for descriptive or quantitative evidence of moderate AS. Follow‐up extended until January 2022. Clinical data were extracted following manual chart review. Socioeconomic variables were collected based on zip‐code–aggregated United States census data. The endpoints were progression from moderate to severe valvular AS, all‐cause mortality, all‐cause hospitalization, and heart failure (HF) hospitalization. The final multivariable model was selected using a variable selection algorithm inspired by greedy causal discovery algorithms.ResultsA total of 34 450 echocardiograms (N = 25 204 patients) were screened during the inclusion period; 367 patients met inclusion criteria and were included in the final analysis. Progression to severe AS was noted in 172 patients (median time to progression 16 months). The final predictive models after variable selection exhibited modest predictive power: progression to severe AS, AUC = 0.68; all‐cause mortality, 0.797; all‐cause hospitalization, 0.629; heart failure hospitalization, 0.744. Variables predictive for the endpoints included comorbidities, echocardiographic variables, and demographics.ConclusionOur findings support further work and exploration of a paradigm shift in the assessment and management of AS, moving beyond traditional measures to a multiparametric model incorporating a broader spectrum of clinical, echocardiographic, and socioeconomic variables.

The telecommunications industry is navigating a transformative era driven by the complexity of 5G, Internet of Things (IoT), edge computing, and software-defined networking. These technologies demand rapid, reliable, and scalable network operations to meet stringent performance, reliability, and regulatory requirements. This paper presents an in-depth study on integrating Artificial Intelligence (AI) into DevOps pipelines to address these challenges. By combining predictive analytics, machine learning (ML), and automation, our proposed AI-Driven DevOps model enhances network agility, optimizes resource allocation, minimizes service disruptions, and enables autonomous operations. We explore practical applications through a detailed case study, address implementation and scalability challenges, propose a conceptual pipeline with mathematical foundations, quantify hypothetical improvements with enhanced metrics, and discuss ethical considerations. The model bridges predictive insights with continuous delivery, fostering resilient, scalable, and intelligent telecom networks capable of adapting to dynamic demands.

We study the emerging collective states in a simple mechanical model of a dense group of self-propelled polar disks with off-centered rotation, confined within a circular arena. Each disk presents self-alignment towards the sum of contact forces acting on it, resulting from disk-substrate interactions, while also displaying mutual alignment with neighbors due to having its center of rotation located a distance RR behind its centroid, so that central contact forces can also introduce torques. The effect of both alignment mechanisms produces a variety of collective states that combine high-frequency localized circular oscillations with low-frequency milling around the center of the arena, in fluid or solid regimes. We consider cases with small/large RR values, isotropic/anisotropic disk-substrate damping, smooth/rough arena boundaries, different densities, and multiple systems sizes, showing that the emergent collective states that we identify are robust, generic, and potentially observable in real-world natural or artificial systems.

e23311 Background: Real-world oncology data integrates structured and unstructured EHR-based information relating to clinical characteristics, treatment patterns, and outcomes for patients. At Tempus, unstructured records are abstracted into structured fields through a uniform, rules-based, human curation process. We aim to measure the performance of our abstraction process by evaluating inter-abstractor reliability, accuracy compared to an independent oncologist, and the utility of Tempus (de-identified) abstracted data for estimating real-world outcomes. Methods: Two randomly selected abstractors (blinded to study participation) independently abstracted the unstructured records of 222 advanced or metastatic non-small cell lung cancer patients (a/mNSCLC). Clinical variables were assessed in the demographic, diagnosis, third-party lab biomarker results, first line treatment (1L), and outcome data domains. A subset of 40 patients were reviewed by an oncologist. The primary measure of inter-abstractor reliability was Gwet’s agreement coefficient (AC). Categorical variables were assessed excluding and including missing data as a category for agreement. Date agreement was calculated for presence/absence, as well as exact, within ±15 days, and ±30 days. The Kaplan-Meier estimate of real-world progression free survival (rwPFS) on combination 1L platinum-based chemotherapy (PBC) and immunotherapy (IO) was derived in 2,980 a/mNSCLC patients diagnosed between 2018-2023. Results: Gwet’s AC was high (≥0.82) between abstractors across demographic, diagnosis, biomarker, and treatment domains. Among the 181 patients where abstractors agreed on 1L class and initiation date within ±30 days, the agreement in progression presence and date was 0.83-0.93. Gwet’s AC was 0.96-1 for death presence and date. Percent agreement was high ranging from 85%-100% between at least one abstractor and the oncologist among categorical variables and 80%-100% within ±30 days for date variables. Median rwPFS on 1L PBC and IO was 7.9 months in line with KEYNOTE-189 and -407. All patients with progression and a non-missing date of progression had a clinically relevant downstream event, 97% with 1L treatment end date, 100% with 2L treatment start date, and 69% with a deceased date. Conclusions: These results demonstrate that the rules-based, human abstraction process as designed is reliable and accurate across the data domains commonly used in insight generation. The resulting data product has utility for estimating real-world outcomes. Domain AC (Min-Max) Demographic (birth date, sex, race, ethnicity, smoking status) 0.96-1 Diagnosis (stage, histology, year of diagnosis) 0.87-0.99 Biomarker (EGFR, ALK, ROS1, PD-L1, BRAF, RET, NTRK) 0.87-1 Treatment (agents, class, dates) 0.82-0.97 Outcomes (progression presence and date) 0.83-0.93 Outcomes (death presence and date) 0.96-1

1041 Background: The three antibody-drug conjugates (ADCs) — sacituzumab govitecan (SG), trastuzumab deruxtecan (T-DXd), and trastuzumab emtansine (T-DM1) — that are FDA-approved for treatment of metastatic breast cancer (MBC) have markedly improved patient outcomes. However, most patients with MBC treated with ADCs ultimately have disease progression via either primary or acquired ADC resistance. Here, we characterized the transcriptomic profile of drug efflux genes in MBC prior to ADC treatment (tx) to elucidate biomarkers of response and resistance to SG, T-Dxd, and T-DM1. Methods: We analyzed the transcriptomic tumor profile of six drug efflux pump genes ( ABCB1 , ABCC1 - 4 , ABCG2 ) generated from pre-tx biopsies collected from patients with MBC (N = 453; 36% TNBC, 26% HR+/HER2-, 20.5% HER2+, 19% NOS) 1 year prior to or up to 15 days post-tx with SG (n = 204), T-DXd (n = 178), or T-DM1 (n = 71). RNA-sequencing data were generated and processed with the Tempus xR assay. The correlation between duration of treatment (DoT) and gene expression was tested for all genes of interest using Pearson’s correlation coefficient. Cox proportional hazards models with risk set adjustment were used to test for associations between pre-tx gene expression and overall survival (OS), where gene expression was modeled as a continuous linear predictor. The proportional hazards assumption for OS was tested, and Cox modeling results were omitted when evidence of non-proportional hazards was detected. Given the exploratory nature of the analyses, all p-values are uncorrected and nominal statistical significance was set at p < 0.05. Results: This diverse cohort had a median age of 52 and a range of races (55% White, 14% Black, 7.1% Other, 24% Unknown). Median DoT across all patients was 130 days. Higher expression of drug efflux pump genes was associated with shorter DoT for T-DXd ( ABCB1 : -0.290, p = 0.017; ABCC1 : -0.274, p = 0.025). Additionally, higher expression of ABCB1 was associated with worse OS for T-DXd (HR: 1.30, 95% CI: 1.10 - 1.53, p = 0.002). In the SG cohort, no significant associations between efflux pump expression and DoT were found, but higher pre-tx ABCC1 and ABCC4 gene expression was associated with worse OS (HR: 1.34, 95% CI: 1.02-1.75, p = 0.034; HR:1.19, 95%CI: 1.00-1.41, p = 0.042). In the T-DM1 cohort, no significant associations were found between efflux pump gene expression and DoT or OS. Conclusions: Multi-modal analysis identified drug efflux pump gene expression as a potential biomarker of resistance, primarily to T-DXd. These findings should be further validated, and combinatorial clinical trial strategies may be explored.