Liquid Biopsy of Lung Cancer Before Pathological Diagnosis Is Associated With Shorter Time to Treatment.

PURPOSE: Studies have investigated the early use of liquid biopsy (LBx) during the diagnostic workup of patients presenting with clinical evidence of advanced lung cancer, but real-world adoption and impact has not been characterized. The aim of this study was to determine whether the use of LBx before diagnosis (Dx; LBx-Dx) enables timely comprehensive genomic profiling (CGP) and shortens time until treatment initiation for advanced non-small-cell lung cancer (aNSCLC). MATERIALS AND METHODS: This study used the Flatiron Health-Foundation Medicine electronic health record-derived deidentified clinicogenomic database of patients with aNSCLC from approximately 280 US cancer clinics. RESULTS: Of 1,076 patients with LBx CGP ordered within 30 days prediagnosis/postdiagnosis, we focused on 56 (5.2%) patients who ordered LBx before diagnosis date (median 8 days between order and diagnosis, range, 1-28). Compared with 1,020 patients who ordered LBx after diagnosis (Dx-LBx), LBx-Dx patients had similar stage and ctDNA tumor fraction (TF). LBx-Dx patients received CGP results a median of 1 day after Dx versus 25 days for Dx-LBx patients. Forty-three percent of LBx-Dx were positive for an National Comprehensive Cancer Network driver, and 32% had ctDNA TF >1% but were driver negative (presumed true negatives). In 748 patients with previously untreated aNSCLC, median time from Dx to therapy was shorter in the LBx-Dx versus Dx-LBx group (21 v 35 days; P < .001). CONCLUSION: Early LBx in anticipation of pathologic diagnosis of aNSCLC was uncommon in this real-world cohort, yet this emerging paradigm was associated with an abbreviated time to CGP results and faster therapy initiation. Forthcoming prospective studies will clarify the utility of LBx in parallel with biopsy for diagnostic confirmation for patients presenting with suspected advanced lung cancer.

Racial and Ethnic Inequities at the Practice and Physician Levels in Timely Next-Generation Sequencing for Patients With Advanced Non-Small-Cell Lung Cancer Treated in the US Community Setting.

PURPOSE: Racial/ethnic inequities in next-generation sequencing (NGS) were examined for patients with advanced non-small-cell lung cancer (aNSCLC) at the practice and physician levels to inform policies to improve equitable quality of care. METHODS: This retrospective study used a nationwide electronic health record-derived deidentified database for patients with aNSCLC diagnosed between April 2018 and March 2022 in the community setting. Timely NGS was an NGS result between initial diagnosis and ≤60 days after advanced diagnosis. We studied how inequities were driven by (1) non-Latinx Black (Black) and Latinx patient under-representation at high testing practices versus (2) Black and Latinx patients being tested at lower rates than non-Latinx White (White) patients, even at the same practice. We defined these two concepts as across inequity and within inequity, respectively, with total inequity as their summation. Mean percentage point inequities were estimated using a Bayesian approach. RESULTS: A total of 12,045 patients (9,981 White; 1,528 Black; 536 Latinx) met study criteria. At the practice level, versus White patients, the mean percentage point difference in NGS testing total inequity was 7.49 for Black and 8.26 for Latinx. Within- and across-practice inequities contributed to total inequity in NGS testing for Black (48% v 52%) and Latinx patients (60% v 40%). At the physician level, versus White patients, the mean percentage point difference in total inequity was 7.73 for Black and 8.81 for Latinx patients. Within- versus across-physician inequities contributed to total inequity for Black and Latinx patients (77% v 23% and 67% v 33%). CONCLUSION: Within-practice, across-practice, and across-physician inequities were main contributors to total inequity in NGS testing, requiring a suite of interventions to effectively address inequities.

Multi-state survival models with treatment effects and biomarkers: Simulations for study design assessment.

BACKGROUND: Comparative effectiveness studies of cancer therapeutics in observational data face confounding by patterns of clinical treatment over time. The validity of survival analysis in longitudinal health records depends on study design choices including index date definition and model specification for covariate adjustment. METHODS: Overall survival in cancer is a multi-state transition process with mortality and treatment switching as competing risks. Parametric Weibull regression quantifies proportionality of hazards across lines of therapy in real-world cohorts of 12 solid tumor types. Study design assessments compare alternative analytic models in simulations with realistic disproportionality. The multi-state simulation framework is adaptable to alternative treatment effect profiles and exposure patterns. RESULTS: Event-specific hazards of treatment-switching and death are not proportional across lines of therapy in 12 solid tumor types. Study designs that include all eligible lines of therapy per subject showed lower bias and variance than designs that select one line per subject. Confounding by line number was effectively mitigated across a range of simulation scenarios by Cox proportional hazards models with stratified baseline hazards and inverse probability of treatment weighting. CONCLUSION: Quantitative study design assessment can inform the planning of observational research in clinical oncology by demonstrating the potential impact of model misspecification. Use of empirical parameter estimates in simulation designs adapts analytic recommendations to the clinical population of interest.

The design and evaluation of hybrid controlled trials that leverage external data and randomization.

Patient-level data from completed clinical studies or electronic health records can be used in the design and analysis of clinical trials. However, these external data can bias the evaluation of the experimental treatment when the statistical design does not appropriately account for potential confounders. In this work, we introduce a hybrid clinical trial design that combines the use of external control datasets and randomization to experimental and control arms, with the aim of producing efficient inference on the experimental treatment effects. Our analysis of the hybrid trial design includes scenarios where the distributions of measured and unmeasured prognostic patient characteristics differ across studies. Using simulations and datasets from clinical studies in extensive-stage small cell lung cancer and glioblastoma, we illustrate the potential advantages of hybrid trial designs compared to externally controlled trials and randomized trial designs.

Deletions on 9p21 are associated with worse outcomes after anti-PD-1/PD-L1 monotherapy but not chemoimmunotherapy.

NCCN guidelines for first-line treatment of advanced non-squamous non-small-cell lung cancer (NSCLC) patients without targetable driver alterations includes either immunotherapy alone or in combination with chemotherapy. In this study, we investigated genomic predictors of survival after immunotherapy to guide this treatment decision. Cox proportional hazards regression was used to identify genomic correlates of survival in a cohort of EGFR/ALK-, non-squamous NSCLC patients treated with first-line pembrolizumab monotherapy (mono-IO) or pembrolizumab in combination with carboplatin/cisplatin and pemetrexed (chemo-IO) within a real-world clinico-genomic database. The effect of deletions on 9p21 was further evaluated in five additional tumor types. Among mono-IO treated non-squamous NSCLC patients, tumors with 9p21.3 gene deletions (CDKN2A, CDKN2B, MTAP) were associated with worse survival compared to the corresponding deletion-negative tumors (CDKN2A deletion HR = 1.8, P = 0.001). However, this association was not observed among chemo-IO treated patients (CDKN2A deletion HR = 1.1, P = 0.4). This finding remained after adjusting for clinical and genomic features including TMB and PD-L1. Deletions at 9p21.3 were not associated with differences in TMB, PD-L1, or tumor inflammation. Due to the high incidence of 9p21.3 deletions across tumor types, we performed a pan-cancer analysis and found CDKN2A deletion-positive tumors had worse survival following first-line immunotherapy treatment in multiple tumor types (HR = 1.4, P < 0.001). These results indicate deletions at 9p21.3 are a putative negative predictor of clinical benefit from first-line immune checkpoint inhibitors and may have utility in choosing between mono-IO vs chemo-IO regimens in NSCLC.

Association of CD274 (PD-L1) Copy Number Changes with Immune Checkpoint Inhibitor Clinical Benefit in Non-Squamous Non-Small Cell Lung Cancer.

BACKGROUND: We sought to characterize response to immune checkpoint inhibitor (ICI) in non-squamous non-small cell lung cancer (NSCLC) across various CD274 copy number gain and loss thresholds and identify an optimal cutoff. MATERIALS AND METHODS: A de-identified nationwide (US) real-world clinico-genomic database was leveraged to study 621 non-squamous NSCLC patients treated with ICI. All patients received second-line ICI monotherapy and underwent comprehensive genomic profiling as part of routine clinical care. Overall survival (OS) from start of ICI, for CD274 copy number gain and loss cohorts across varying copy number thresholds, were assessed. RESULTS: Among the 621 patients, patients with a CD274 CN greater than or equal to specimen ploidy +2 (N = 29) had a significantly higher median (m) OS when compared with the rest of the cohort (N = 592; 16.1 [8.9-37.3] vs 8.6 [7.1-10.9] months, hazard ratio (HR) = 0.6 [0.4-1.0], P-value = .05). Patients with a CD274 copy number less than specimen ploidy (N = 299) trended toward a lower mOS when compared to the rest of the cohort (N = 322; 7.5 [5.9-11.3] vs 9.6 [7.9-12.8] months, HR = 0.9 [0.7-1.1], P-value = .3). CONCLUSION: This work shows that CD274 copy number gains at varying thresholds predict different response to ICI blockade in non-squamous NSCLC. Considering these data, prospective clinical trials should further validate these findings, specifically in the context of PD-L1 IHC test results.

Bayesian data fusion: Probabilistic sensitivity analysis for unmeasured confounding using informative priors based on secondary data.

Bayesian causal inference offers a principled approach to policy evaluation of proposed interventions on mediators or time-varying exposures. Building on the Bayesian g-formula method introduced by Keil et al., we outline a general approach for the estimation of population-level causal quantities involving dynamic and stochastic treatment regimes, including regimes related to mediation estimands such as natural direct and indirect effects. We further extend this approach to propose a Bayesian data fusion (BDF), an algorithm for performing probabilistic sensitivity analysis when a confounder unmeasured in a primary data set is available in an external data source. When the relevant relationships are causally transportable between the two source populations, BDF corrects confounding bias and supports causal inference and decision-making within the main study population without sharing of the individual-level external data set. We present results from a simulation study comparing BDF to two common frequentist correction methods for unmeasured mediator-outcome confounding bias in the mediation setting. We use these methods to analyze data on the role of stage at cancer diagnosis in contributing to Black-White colorectal cancer survival disparities.

The use of external control data for predictions and futility interim analyses in clinical trials.

BACKGROUND: External control (EC) data from completed clinical trials and electronic health records can be valuable for the design and analysis of future clinical trials. We discuss the use of EC data for early stopping decisions in randomized clinical trials (RCTs). METHODS: We specify interim analyses (IAs) approaches for RCTs, which allow investigators to integrate external data into early futility stopping decisions. IAs utilize predictions based on early data from the RCT, possibly combined with external data. These predictions at IAs express the probability that the trial will generate significant evidence of positive treatment effects. The trial is discontinued if this predictive probability becomes smaller than a prespecified threshold. We quantify efficiency gains and risks associated with the integration of external data into interim decisions. We then analyze a collection of glioblastoma (GBM) data sets, to investigate if the balance of efficiency gains and risks justify the integration of external data into the IAs of future GBM RCTs. RESULTS: Our analyses illustrate the importance of accounting for potential differences between the distributions of prognostic variables in the RCT and in the external data to effectively leverage external data for interim decisions. Using GBM data sets, we estimate that the integration of external data increases the probability of early stopping of ineffective experimental treatments by up to 25% compared to IAs that do not leverage external data. Additionally, we observe a reduction of the probability of early discontinuation for effective experimental treatments, which improves the RCT power. CONCLUSION: Leveraging external data for IAs in RCTs can support early stopping decisions and reduce the number of enrolled patients when the experimental treatment is ineffective.

Association and interplay of surgeon and hospital volume with mortality after open abdominal aortic aneurysm repair in the modern era.

OBJECTIVE: Operative volume has been used as a marker of quality. Research from previous decades has suggested minimum open abdominal aortic aneurysm (AAA) repair volume requirements for surgeons of 9 to 13 open AAA repairs annually and for hospitals of 18 open AAA repairs annually to purportedly achieve acceptable results. Given concerns regarding the decreased frequency of open repairs in the endovascular era, we examined the association of surgeon and hospital volume with the 30- and 90-day mortality in the Vascular Quality Initiative (VQI) registry. METHODS: Patients who had undergone elective open AAA repair from 2013 to 2018 were identified in the VQI registry. We performed a cross-sectional evaluation of the association between the average hospital and surgeon volume and 30-day postoperative mortality using a hierarchical Bayesian model. Cross-level interactions were permitted, and random surgeon- and hospital-level intercepts were used to account for clustering. The mortality results were adjusted by standardizing to the observed distribution of relevant covariates in the overall cohort. The outcomes were compared to the Society for Vascular Surgery guidelines recommended criteria of <5% perioperative mortality. RESULTS: A total of 3078 patients had undergone elective open AAA repair by 520 surgeons at 128 hospitals. The 30- and 90-day risks of postoperative mortality were 4.1% (n = 126) and 5.4% (n = 166), respectively. The mean surgeon volume and hospital volume both correlated inversely with the 30-day mortality. Averaged across all patients and hospitals, we found a 96% probability that surgeons who performed an average of four or more repairs per year achieved <5% 30-day mortality. Substantial interplay was present between surgeon volume and hospital volume. For example, at lower volume hospitals performing an average of five repairs annually, <5% 30-day mortality would be expected 69% of the time for surgeons performing an average of three operations annually. In contrast, at higher volume hospitals performing an average of 40 repairs annually, a <5% 30-day mortality would be expected 96% of the time for surgeons performing an average of three operations annually. As hospital volume increased, a diminishing difference occurred in 30-day mortality between lower and higher volume surgeons. Likewise, as surgeon volume increased, a diminishing difference was found in 30-day mortality between the lower and higher volume hospitals. CONCLUSIONS: Surgeons and hospitals in the VQI registry achieved mortality outcomes of <5% (Society for Vascular Surgery guidelines), with an average surgeon volume that was substantially lower compared with previous reports. Furthermore, when considering the development of minimal surgeon volume guidelines, it is important to contextualize the outcomes within the hospital volumes.

Somatic HLA Class I Loss Is a Widespread Mechanism of Immune Evasion Which Refines the Use of Tumor Mutational Burden as a Biomarker of Checkpoint Inhibitor Response.

Neoantigen presentation arises as a result of tumor-specific mutations and is a critical component of immune surveillance that can be abrogated by somatic LOH of the human leukocyte antigen class I (HLA-I) locus. To understand the role of HLA-I LOH in oncogenesis and treatment, we utilized a pan-cancer genomic dataset of 83,644 patient samples, a small subset of which had treatment outcomes with immune checkpoint inhibitors (ICI). HLA-I LOH was common (17%) and unexpectedly had a nonlinear relationship with tumor mutational burden (TMB). HLA-I LOH was frequent at intermediate TMB, yet prevalence decreased above 30 mutations/megabase, suggesting highly mutated tumors require alternate immune evasion mechanisms. In ICI-treated patients with nonsquamous non-small cell lung cancer, HLA-I LOH was a significant negative predictor of overall survival. Survival prediction improved when combined with TMB, suggesting TMB with HLA-I LOH may better identify patients likely to benefit from ICIs. SIGNIFICANCE: This work shows the pan-cancer landscape of HLA-I LOH, revealing an unexpected "Goldilocks" relationship between HLA-I LOH and TMB, and demonstrates HLA-I LOH as a significant negative predictor of outcomes after ICI treatment. These data informed a combined predictor of outcomes after ICI and have implications for tumor vaccine development.This article is highlighted in the In This Issue feature, p. 211.

Functional Dependence and Mortality in the International Dialysis Outcomes and Practice Patterns Study (DOPPS).

BACKGROUND: Patients receiving long-term dialysis have among the highest mortality and hospitalization rates. In the nonrenal literature, functional dependence is recognized as a contributor to subsequent disability, recurrent hospitalization, and increased mortality. A higher burden of functional dependence with progressive worsening of kidney function has been observed in several studies, suggesting that functional dependence may contribute to both morbidity and mortality in dialysis patients. STUDY DESIGN: Prospective cohort study. SETTING & PARTICIPANTS: 7,226 hemodialysis patients from 12 countries in the DOPPS (Dialysis Outcomes and Practice Patterns Study) phase 4 (2009-2011) with self-reported data for functional status. PREDICTOR: Patients' ability to perform 13 basic and instrumental activities of daily living was summarized to create an overall functional status score (range, 1.25 [most dependent] to 13 [functionally independent]). OUTCOME: Cox regression was used to estimate the association between functional status and all-cause mortality, adjusting for several demographic and clinical risk factors for mortality. Median follow-up was 17.2 months. RESULTS: The proportion of patients who could perform each activity of daily living task without assistance ranged from 97% (eating) to 47% (doing housework). 36% of patients could perform all 13 tasks without assistance (functional status = 13), and 14% of patients had high functional dependence (functional status < 8). Functionally independent patients were younger and had many indicators of better health status, including higher quality of life. Compared with functionally independent patients, the adjusted HR for mortality was 2.37 (95% CI, 1.92-2.94) for patients with functional status < 8. LIMITATIONS: Possible nonresponse bias and residual confounding. CONCLUSIONS: We found a high burden of functional dependence across all age groups and across all DOPPS countries. When adjusting for several known mortality risk factors, including age, access type, cachexia, and multimorbidity, functional dependence was a strong consistent predictor of mortality.

Does cannabis onset trigger cocaine onset? A case-crossover approach.

Psychiatric researchers tend to select the discordant co-twin design when they seek to hold constant genetic influence while estimating exposure-associated disease risk. The epidemiologic case-crossover research design developed for the past two decades represents a viable alternative, not often seen in psychiatric studies. Here, we turn to the epidemiologic case-crossover approach to examine the idea that cannabis onset is a proximal trigger for cocaine use, with the power of "subject-as-own-control" research used to hold constant antecedent characteristics of the individual drug user, including genetic influence and other traits experienced up to the time of the observed hazard and control intervals. Data are from newly incident cocaine users identified in the 2002-2006 US National Surveys on Drug Use and Health. Among these cocaine users, 48 had both cannabis onset and cocaine onset in the same month-long hazard interval; the expected value is 30 users, based on the control interval we had pre-specified for case-crossover estimation (estimated relative risk, RR = 1.6; exact mid-p = 0.042). Within the framework of a subject-as-own-control design, the evidence is consistent with the hypothesis that cannabis onset is a proximal trigger for cocaine use, with genetic influences (and many environmental conditions and processes) held constant. Limitations are noted and implications are discussed.