Proportional Hazards Violations in Phase 3 Cancer Clinical Trials: A Potential Source of Trial Misinterpretation.

BACKGROUND: Survival analyses of novel agents with long-term responders often exhibit differential hazard rates over time. Such proportional hazards violations (PHVs) may reduce the power of the log-rank test and lead to misinterpretation of trial results. We aimed to characterize the incidence and study attributes associated with PHVs in phase 3 oncology trials and assess the utility of restricted mean survival time (RMST) and MaxCombo as additional analyses. METHODS: Clinicaltrials.gov and PubMed were searched to identify 2-arm, randomized, phase 3 superiority-design cancer trials with time-to-event primary endpoints and published results through 2020. Patient-level data were reconstructed from published Kaplan-Meier curves. PHVs were assessed using Schoenfeld residuals. RESULTS: Three hundred fifty-seven Kaplan-Meier comparisons across 341 trials were analyzed, encompassing 292,831 enrolled patients. PHVs were identified in 85/357 (23.8%; 95%CI 19.7%, 28.5%) comparisons. In multivariable analysis, non-OS endpoints (odds ratio [OR] 2.16 [95%CI 1.21, 3.87]; P=.009) were associated with higher odds of PHVs, and immunotherapy comparisons (OR 1.94 [95%CI 0.98, 3.86]; P=.058) were weakly suggestive of higher odds of PHVs. Few trials with PHVs (25/85, 29.4%) pre-specified a statistical plan to account for PHVs. Fourteen trials with PHVs exhibited discordant statistical signals with RMST or MaxCombo, of which ten (71%) reported negative results. CONCLUSION: PHVs are common across therapy types, and attempts to account for PHVs in statistical design are lacking despite the potential for results exhibiting non-proportional hazards to be misinterpreted.

Signet ring cells and conditional survival after trimodality therapy for esophageal adenocarcinoma.

BACKGROUND AND METHODS: Although signet ring cell (SRC) histology is associated with resistance to neoadjuvant chemoradiotherapy and worse overall survival (OS) in esophageal adenocarcinoma (EAC), its prognostic relationship among patients who survive the early period following resection is unknown. EAC patients who underwent trimodality therapy at a single institution (2006-2018) were identified. Bayesian multivariable regression (BMR) analyses of OS and additional OS from a 3-year landmark were performed. RESULTS: Of 631 patients, SRCs were present in 16.0% (N = 101). SRC was associated with shorter median OS (45.8 [95% confidence interval: 31.0-96.7] vs. 79.8 [63.0-107.2] months; p = 0.014). In BMR analysis, the absence of an SRC component was moderately associated with improved OS (probability of beneficial effect, PBE = 0.879). Three-year conditional BMR analysis of additional OS (N = 357) showed that SRC status no longer had a prognostic effect (PBE = 0.546); higher pathological stage was strongly associated with worse additional OS (PBE < 0.001). CONCLUSIONS: The presence of SRC portends worse OS following trimodality therapy for EAC. However, this prognostic impact is dynamic and abates by 3 years postoperatively. In contrast, a higher pathological stage is strongly associated with poor overall and 3-year conditional survival. DISCUSSION: These findings may inform postoperative patient counseling and surveillance protocols.

Bayesian sequential monitoring strategies for trials of digestive cancer therapeutics.

BACKGROUND: New therapeutics in oncology have presented challenges to existing paradigms and trial designs in all phases of drug development. As a motivating example, we considered an ongoing phase II trial planned to evaluate the combination of a MET inhibitor and an anti-PD-L1 immunotherapy to treat advanced oesogastric carcinoma. The objective of the paper was to exemplify the planning of an adaptive phase II trial with novel anti-cancer agents, including prolonged observation windows and joint sequential evaluation of efficacy and toxicity. METHODS: We considered various candidate designs and computed decision rules assuming correlations between efficacy and toxicity. Simulations were conducted to evaluate the operating characteristics of all designs. RESULTS: Design approaches allowing continuous accrual, such as the time-to-event Bayesian Optimal Phase II design (TOP), showed good operating characteristics while ensuring a reduced trial duration. All designs were sensitive to the specification of the correlation between efficacy and toxicity during planning, but TOP can take that correlation into account more easily. CONCLUSIONS: While specifying design working hypotheses requires caution, Bayesian approaches such as the TOP design had desirable operating characteristics and allowed incorporating concomittant information, such as toxicity data from concomitant observations in another relevant patient population (e.g., defined by mutational status).

Development and clinical implementation of a digital workflow utilizing 3D-printed oral stents for patients with head and neck cancer receiving radiotherapy.

OBJECTIVES: We describe the development of 3D-printed stents using our digital workflow and their effects on patients enrolled in the lead-in phase of a multi-center, randomized Phase-II trial. MATERIALS AND METHODS: Digital dental models were created for patients using intraoral scanning. Digital processes were implemented to develop the mouth-opening, tongue-depressing, and tongue-lateralizing stents using stereolithography. Time spent and material 3D-printing costs were measured. Physicians assessed mucositis using the Oral Mucositis Assessment Scale (OMAS) and collected MD Anderson Symptom Inventory (MDASI) reports and adverse events (AEs) from patients at various time points (TPs). OMAS and MDASI results were evaluated using paired t-test analysis. RESULTS: 18 patients enrolled into the lead-in phase across 6 independent clinical sites in the USA. 15 patients received stents (average design and fabrication time, 8 h; average material 3D-printing cost, 11 USD). 10 eligible patients with complete OMAS and MDASI reports across all TPs were assessed. OMAS increased significantly from baseline to week 3 of treatment (mean difference = 0.34; 95 % CI, 0.09-0.60; p = 0.01). MDASI increased significantly from baseline to week 3 of treatment (mean difference = 1.02; 95 % CI, 0.40-1.70; p = 0.005), and week 3 of treatment to end of treatment (mean difference = 1.90; 95 % CI, 0.90-2.92; p = 0.002). AEs (grades 1-3) were reported by patients across TPs. Mucositis and radiation dermatitis were primarily attributed to chemoradiation. CONCLUSIONS: 3D-printed stents were successfully fabricated and well tolerated by patients. As patients enroll in the randomized phase of this trial, data herein will establish a baseline for comparative analysis.

REDOMA: Bayesian random-effects dose-optimization meta-analysis using spike-and-slab priors.

The rise of cutting-edge precision cancer treatments has led to a growing significance of the optimal biological dose (OBD) in modern oncology trials. These trials now prioritize the consideration of both toxicity and efficacy simultaneously when determining the most desirable dosage for treatment. Traditional approaches in early-phase oncology trials have conventionally relied on the assumption of a monotone relationship between treatment efficacy and dosage. However, this assumption may not hold valid for novel oncology therapies. In reality, the dose-efficacy curve of such treatments may reach a plateau at a specific dose, posing challenges for conventional methods in accurately identifying the OBD. Furthermore, achieving reliable identification of the OBD is typically not possible based on a single small-sample trial. With data from multiple phase I and phase I/II trials, we propose a novel Bayesian random-effects dose-optimization meta-analysis (REDOMA) approach to identify the OBD by synthesizing toxicity and efficacy data from each trial. The REDOMA method can address trials with heterogeneous characteristics. We adopt a curve-free approach based on a Gamma process prior to model the average dose-toxicity relationship. In addition, we utilize a Bayesian model selection framework that uses the spike-and-slab prior as an automatic variable selection technique to eliminate monotonic constraints on the dose-efficacy curve. The good performance of the REDOMA method is confirmed by extensive simulation studies.

Comparative review of novel model-assisted designs for phase I/II clinical trials.

In recent years, both model-based and model-assisted designs have emerged to efficiently determine the optimal biological dose (OBD) in phase I/II trials for immunotherapy and targeted cellular agents. Model-based designs necessitate repeated model fitting and computationally intensive posterior sampling for each dose-assignment decision, limiting their practical application in real trials. On the other hand, model-assisted designs employ simple statistical models and facilitate the precalculation of a decision table for use throughout the trial, eliminating the need for repeated model fitting. Due to their simplicity and transparency, model-assisted designs are often preferred in phase I/II trials. In this paper, we systematically evaluate and compare the operating characteristics of several recent model-assisted phase I/II designs, including TEPI, PRINTE, Joint i3+3, BOIN-ET, STEIN, uTPI, and BOIN12, in addition to the well-known model-based EffTox design, using comprehensive numerical simulations. To ensure an unbiased comparison, we generated 10,000 dosing scenarios using a random scenario generation algorithm for each predetermined OBD location. We thoroughly assess various performance metrics, such as the selection percentages, average patient allocation to OBD, and overdose percentages across the eight designs. Based on these assessments, we offer design recommendations tailored to different objectives, sample sizes, and starting dose locations.

Design and sample size determination for multiple-dose randomized phase II trials for dose optimization.

The U.S. Food and Drug Administration (FDA) has launched Project Optimus to shift dose selection from the maximum tolerated dose (MTD) to the dose that produces the optimal risk-benefit tradeoff. One approach highlighted in the FDA's guidance involves conducting a randomized phase II trial following the completion of a phase I trial, where multiple doses (typically including the MTD and one or two doses lower than the MTD) are compared to identify the optimal dose that maximizes the benefit-risk tradeoff. This article focuses on the design of such a multiple-dose randomized trial, specifically the determination of the sample size. We generalized the standard definitions of type I error and power to accommodate the unique characteristics of dose optimization and derived a decision rule along with an algorithm to determine the optimal sample size. The resulting design is referred to as MERIT (Multiple-dosE RandomIzed Trial design for dose optimization based on toxicity and efficacy). Simulation studies demonstrate that MERIT has desirable operating characteristics, and a sample size between 20 and 40 per dosage arm often offers reasonable power and type I errors to ensure patient safety and benefit. To facilitate the implementation of the MERIT design, we provide software, available at https://www.trialdesign.org.

A Bayesian quasi-likelihood design for identifying the minimum effective dose and maximum utility dose in dose-ranging studies.

Most existing dose-ranging study designs focus on assessing the dose-efficacy relationship and identifying the minimum effective dose. There is an increasing interest in optimizing the dose based on the benefit-risk tradeoff. We propose a Bayesian quasi-likelihood dose-ranging design that jointly considers safety and efficacy to simultaneously identify the minimum effective dose and the maximum utility dose to optimize the benefit-risk tradeoff. The binary toxicity endpoint is modeled using a beta-binomial model. The efficacy endpoint is modeled using the quasi-likelihood approach to accommodate various types of data (e.g. binary, ordinal or continuous) without imposing any parametric assumptions on the dose-response curve. Our design utilizes a utility function as a measure of benefit-risk tradeoff and adaptively assign patients to doses based on the doses' likelihood of being the minimum effective dose and maximum utility dose. The design takes a group-sequential approach. At each interim, the doses that are deemed overly toxic or futile are dropped. At the end of the trial, we use posterior probability criteria to assess the strength of the dose-response relationship for establishing the proof-of-concept. If the proof-of-concept is established, we identify the minimum effective dose and maximum utility dose. Our simulation study shows that compared with some existing designs, the Bayesian quasi-likelihood dose-ranging design is robust and yields competitive performance in establishing proof-of-concept and selecting the minimum effective dose. Moreover, it includes an additional feature for further maximum utility dose selection.

Clonal dominance defines metastatic dissemination in pancreatic cancer.

Tumors represent ecosystems where subclones compete during tumor growth. While extensively investigated, a comprehensive picture of the interplay of clonal lineages during dissemination is still lacking. Using patient-derived pancreatic cancer cells, we created orthotopically implanted clonal replica tumors to trace clonal dynamics of unperturbed tumor expansion and dissemination. This model revealed the multifaceted nature of tumor growth, with rapid changes in clonal fitness leading to continuous reshuffling of tumor architecture and alternating clonal dominance as a distinct feature of cancer growth. Regarding dissemination, a large fraction of tumor lineages could be found at secondary sites each having distinctive organ growth patterns as well as numerous undescribed behaviors such as abortive colonization. Paired analysis of primary and secondary sites revealed fitness as major contributor to dissemination. From the analysis of pro- and nonmetastatic isogenic subclones, we identified a transcriptomic signature able to identify metastatic cells in human tumors and predict patients' survival.

On the relative conservativeness of Bayesian logistic regression method in oncology dose-finding studies.

The Bayesian logistic regression method (BLRM) is a widely adopted and flexible design for finding the maximum tolerated dose in oncology phase I studies. However, the BLRM design has been criticized in the literature for being overly conservative due to the use of the overdose control rule. Recently, a discussion paper titled "Improving the performance of Bayesian logistic regression model with overall control in oncology dose-finding studies" in Statistics in Medicine has proposed an overall control rule to address the "excessive conservativeness" of the standard BLRM design. In this short communication, we discuss the relative conservativeness of the standard BLRM design and also suggest a dose-switching rule to further enhance its performance.

A generalized calibrated Bayesian hierarchical modeling approach to basket trials with multiple endpoints.

A basket trial simultaneously evaluates a treatment in multiple cancer subtypes, offering an effective way to accelerate drug development in multiple indications. Many basket trials are designed and monitored based on a single efficacy endpoint, primarily the tumor response. For molecular targeted or immunotherapy agents, however, a single efficacy endpoint cannot adequately characterize the treatment effect. It is increasingly important to use more complex endpoints to comprehensively assess the risk-benefit profile of such targeted therapies. We extend the calibrated Bayesian hierarchical modeling approach to monitor phase II basket trials with multiple endpoints. We propose two generalizations, one based on the latent variable approach and the other based on the multinomial-normal hierarchical model, to accommodate different types of endpoints and dependence assumptions regarding information sharing. We introduce shrinkage parameters as functions of statistics measuring homogeneity among subgroups and propose a general calibration approach to determine the functional forms. Theoretical properties of the generalized hierarchical models are investigated. Simulation studies demonstrate that the monitoring procedure based on the generalized approach yields desirable operating characteristics.

Survival outcomes and toxicity of adjuvant immunotherapy after definitive concurrent chemotherapy with proton beam radiation therapy for patients with inoperable locally advanced non-small cell lung carcinoma.

INTRODUCTION: Adjuvant immunotherapy (IO) following concurrent chemotherapy and photon radiation therapy confers an overall survival (OS) benefit for patients with inoperable locally advanced non-small cell lung carcinoma (LA-NSCLC); however, outcomes of adjuvant IO after concurrent chemotherapy with proton beam therapy (CPBT) are unknown. We investigated OS and toxicity after CPBT with adjuvant IO versus CPBT alone for inoperable LA-NSCLC. MATERIALS AND METHODS: We analyzed 354 patients with LA-NSCLC who were prospectively treated with CPBT with or without adjuvant IO from 2009 to 2021. Optimal variable ratio propensity score matching (PSM) matched CPBT with CPBT + IO patients. Survival was estimated with the Kaplan-Meier method and compared with log-rank tests. Multivariable Cox proportional hazards regression evaluated the effect of IO on disease outcomes. RESULTS: Median age was 70 years; 71 (20%) received CPBT + IO and 283 (80%) received CPBT only. After PSM, 71 CPBT patients were matched with 71 CPBT + IO patients. Three-year survival rates for CPBT + IO vs CPBT were: OS 67% vs 30% (P < 0.001) and PFS 59% vs 35% (P = 0.017). Three-year LRFS (P = 0.137) and DMFS (P = 0.086) did not differ. Receipt of adjuvant IO was a strong predictor of OS (HR 0.40, P = 0.001) and PFS (HR 0.56, P = 0.030), but not LRFS (HR 0.61, P = 0.121) or DMFS (HR 0.61, P = 0.136). There was an increased incidence of grade ≥3 esophagitis in the CPBT-only group (6% CPBT + IO vs 17% CPBT, P = 0.037). CONCLUSION: This study, one of the first to investigate CPBT followed by IO for inoperable LA-NSCLC, showed that IO conferred survival benefits with no increased rates of toxicity.

Adaptive Bayesian information borrowing methods for finding and optimizing subgroup-specific doses.

In precision oncology, integrating multiple cancer patient subgroups into a single master protocol allows for the simultaneous assessment of treatment effects in these subgroups and promotes the sharing of information between them, ultimately reducing sample sizes and costs and enhancing scientific validity. However, the safety and efficacy of these therapies may vary across different subgroups, resulting in heterogeneous outcomes. Therefore, identifying subgroup-specific optimal doses in early-phase clinical trials is crucial for the development of future trials. In this article, we review various innovative Bayesian information-borrowing strategies that aim to determine and optimize subgroup-specific doses. Specifically, we discuss Bayesian hierarchical modeling, Bayesian clustering, Bayesian model averaging or selection, pairwise borrowing, and other relevant approaches. By employing these Bayesian information-borrowing methods, investigators can gain a better understanding of the intricate relationships between dose, toxicity, and efficacy in each subgroup. This increased understanding significantly improves the chances of identifying an optimal dose tailored to each specific subgroup. Furthermore, we present several practical recommendations to guide the design of future early-phase oncology trials involving multiple subgroups when using the Bayesian information-borrowing methods.

Association Between Symptom Burden and Early Lymphatic Abnormalities After Regional Nodal Irradiation for Breast Cancer.

PURPOSE: Dermal backflow visualized on near-infrared fluorescence lymphatic imaging (NIRF-LI) signals preclinical lymphedema that precedes the development of volumetrically defined lymphedema. We sought to evaluate whether dermal backflow correlates with patient-reported lymphedema outcomes (PRLO) surveys in breast cancer patients treated with regional nodal irradiation (RNI). METHODS AND MATERIALS: Patients with breast cancer planned for axillary dissection and RNI prospectively underwent perometry, NIRF-LI, and PRLOs (the Lymphedema Symptom Intensity and Distress Survey [LSIDS] and QuickDASH) at baseline, after surgery, and at 6, 12, and 18 months after radiation. Clinical lymphedema was defined as an arm volume increase ≥5% over baseline. Trends over time were assessed using analysis of variance testing. The association between survey responses and both dermal backflow and lymphedema was assessed using a linear mixed-effects model. RESULTS: Sixty participants completed at least 2 sets of measurements and surveys and were eligible for analysis. Fifty-four percent of patients had cT3-T4 disease, 53% cN3 disease, and 75% had a body mass index >25. Dermal backflow and clinical lymphedema increased from 10% to 85% and from 0% to 40%, respectively, from baseline to 18 months. In the adjusted model, soft tissue sensation, neurologic sensation, and functional LSIDS subscale scores were associated with presence of dermal backflow (all P < .05). Both dermal backflow and lymphedema were associated with QuickDASH score (P < .05). CONCLUSIONS: In this high-risk cohort, we found highly prevalent early signs of lymphedema, with increased symptom burden from baseline. Presence of dermal backflow correlated with PRLO measures, highlighting a potential NIRF-LI use to identify patients for early intervention trials after RNI.

Novel Clinical Trial Designs with Dose Optimization to Improve Long-term Outcomes.

Conventional designs for choosing a dose for a new therapy may select doses that are unsafe or ineffective and fail to optimize progression-free survival time, overall survival time, or response/remission duration. We explain and illustrate limitations of conventional dose-finding designs and make four recommendations to address these problems. When feasible, a dose-finding design should account for long-term outcomes, include screening rules that drop unsafe or ineffective doses, enroll an adequate sample size, and randomize patients among doses. As illustrations, we review three designs that include one or more of these features. The first illustration is a trial that randomized patients among two cell therapy doses and standard of care in a setting where it was assumed on biological grounds that dose toxicity and dose-response curves did not necessarily increase with cell dose. The second design generalizes phase I-II by first identifying a set of candidate doses, rather than one dose, randomizing additional patients among the candidates, and selecting an optimal dose to maximize progression-free survival over a longer follow-up period. The third design combines a phase I-II trial and a group sequential randomized phase III trial by using survival time data available after the first stage of phase III to reoptimize the dose selected in phase I-II. By incorporating one or more of the recommended features, these designs improve the likelihood that a selected dose or schedule will be optimal, and thus will benefit future patients and obtain regulatory approval.

Local continual reassessment methods for dose finding and optimization in drug-combination trials.

Due to the limited sample size and large dose exploration space, obtaining a desirable dose combination is a challenging task in the early development of combination treatments for cancer patients. Most existing designs for optimizing the dose combination are model-based, requiring significant efforts to elicit parameters or prior distributions. Model-based designs also rely on intensive model calibration and may yield unstable performance in the case of model misspecification or sparse data. We propose to employ local, underparameterized models for dose exploration to reduce the hurdle of model calibration and enhance the design robustness. Building upon the framework of the partial ordering continual reassessment method, we develop local data-based continual reassessment method designs for identifying the maximum tolerated dose combination, using toxicity only, and the optimal biological dose combination, using both toxicity and efficacy, respectively. The local data-based continual reassessment method designs only model the local data from neighboring dose combinations. Therefore, they are flexible in estimating the local space and circumventing unstable characterization of the entire dose-exploration surface. Our simulation studies show that our approach has competitive performance compared to widely used methods for finding maximum tolerated dose combination, and it has advantages over existing model-based methods for optimizing optimal biological dose combination.

On singular values of large dimensional lag-τ sample auto-correlation matrices.

We study the limiting behavior of singular values of a lag-τ sample auto-correlation matrix Rτϵ of large dimensional vector white noise process, the error term ϵ in the high-dimensional factor model. We establish the limiting spectral distribution (LSD) that characterizes the global spectrum of Rτϵ, and derive the limit of its largest singular value. All the asymptotic results are derived under the high-dimensional asymptotic regime where the data dimension and sample size go to infinity proportionally. Under mild assumptions, we show that the LSD of Rτϵ is the same as that of the lag-τ sample auto-covariance matrix. Based on this asymptotic equivalence, we additionally show that the largest singular value of Rτϵ converges almost surely to the right end point of the support of its LSD. Based on these results, we further propose two estimators of total number of factors with lag-τ sample auto-correlation matrices in a factor model. Our theoretical results are fully supported by numerical experiments as well.

Identification of Key Elements in Prostate Cancer for Ontology Building via a Multidisciplinary Consensus Agreement.

BACKGROUND: Clinical data collection related to prostate cancer (PCa) care is often unstructured or heterogeneous among providers, resulting in a high risk for ambiguity in its meaning when sharing or analyzing data. Ontologies, which are shareable formal (i.e., computable) representations of knowledge, can address these challenges by enabling machine-readable semantic interoperability. The purpose of this study was to identify PCa-specific key data elements (KDEs) for standardization in clinic and research. METHODS: A modified Delphi method using iterative online surveys was performed to report a consensus agreement on KDEs by a multidisciplinary panel of 39 PCa specialists. Data elements were divided into three themes in PCa and included (1) treatment-related toxicities (TRT), (2) patient-reported outcome measures (PROM), and (3) disease control metrics (DCM). RESULTS: The panel reached consensus on a thirty-item, two-tiered list of KDEs focusing mainly on urinary and rectal symptoms. The Expanded Prostate Cancer Index Composite (EPIC-26) questionnaire was considered most robust for PROM multi-domain monitoring, and granular KDEs were defined for DCM. CONCLUSIONS: This expert consensus on PCa-specific KDEs has served as a foundation for a professional society-endorsed, publicly available operational ontology developed by the American Association of Physicists in Medicine (AAPM) Big Data Sub Committee (BDSC).

A novel statistical method for decontaminating T-cell receptor sequencing data.

The T-cell receptor (TCR) repertoire is highly diverse among the population and plays an essential role in initiating multiple immune processes. TCR sequencing (TCR-seq) has been developed to profile the T cell repertoire. Similar to other high-throughput experiments, contamination can happen during several steps of TCR-seq, including sample collection, preparation and sequencing. Such contamination creates artifacts in the data, leading to inaccurate or even biased results. Most existing methods assume 'clean' TCR-seq data as the starting point with no ability to handle data contamination. Here, we develop a novel statistical model to systematically detect and remove contamination in TCR-seq data. We summarize the observed contamination into two sources, pairwise and cross-cohort. For both sources, we provide visualizations and summary statistics to help users assess the severity of the contamination. Incorporating prior information from 14 existing TCR-seq datasets with minimum contamination, we develop a straightforward Bayesian model to statistically identify contaminated samples. We further provide strategies for removing the impacted sequences to allow for downstream analysis, thus avoiding any need to repeat experiments. Our proposed model shows robustness in contamination detection compared with a few off-the-shelf detection methods in simulation studies. We illustrate the use of our proposed method on two TCR-seq datasets generated locally.