Log-Rank Test vs MaxCombo and Difference in Restricted Mean Survival Time Tests for Comparing Survival Under Nonproportional Hazards in Immuno-oncology Trials: A Systematic Review and Meta-analysis

Biomarker-guided adaptive enrichment design with threshold detection for clinical trials with time-to-event outcome

Biomarker-guided designs are increasingly used to evaluate personalized treatments based on patients' biomarker status in Phase II and III clinical trials. With adaptive enrichment, these designs can improve the efficiency of evaluating the treatment effect in biomarker-positive patients by increasing their proportion in the randomized trial. While time-to-event outcomes are often used as the primary endpoint to measure treatment effects for a new therapy in severe diseases like cancer and cardiovascular diseases, there is limited research on biomarker-guided adaptive enrichment trials in this context. Such trials almost always adopt hazard ratio methods for statistical measurement of treatment effects. In contrast, restricted mean survival time (RMST) has gained popularity for analyzing time-to-event outcomes because it offers more straightforward interpretations of treatment effects and does not require the proportional hazard assumption. This paper proposes a two-stage biomarker-guided adaptive RMST design with threshold detection and patient enrichment. We develop sophisticated methods for identifying the optimal biomarker threshold and biomarker-positive subgroup, treatment effect estimators, and approaches for type I error rate, power analysis, and sample size calculation. We present a numerical example of re-designing an oncology trial. An extensive simulation study is conducted to evaluate the performance of the proposed design.

Society for Immunotherapy of Cancer (SITC) consensus statement on essential biomarkers for immunotherapy clinical protocols

Immunotherapy of cancer is now an essential pillar of treatment for patients with many individual tumor types. Novel immune targets and technical advances are driving a rapid exploration of new treatment strategies incorporating immune agents in cancer clinical practice. Immunotherapies perturb a complex system of interactions among genomically unstable tumor cells, diverse cells within the tumor microenvironment including the systemic adaptive and innate immune cells. The drive to develop increasingly effective immunotherapy regimens is tempered by the risk of immune-related adverse events. Evidence-based biomarkers that measure the potential for therapeutic response and/or toxicity are critical to guide optimal patient care and contextualize the results of immunotherapy clinical trials. Responding to the lack of guidance on biomarker testing in early-phase immunotherapy clinical trials, we propose a definition and listing of essential biomarkers recommended for inclusion in all such protocols. These recommendations are based on consensus provided by the Society for Immunotherapy of Cancer (SITC) Clinical Immuno-Oncology Network (SCION) faculty with input from the SITC Pathology and Biomarker Committees and the Journal for ImmunoTherapy of Cancer readership. A consensus-based selection of essential biomarkers was conducted using a Delphi survey of SCION faculty. Regular updates to these recommendations are planned. The inaugural list of essential biomarkers includes complete blood count with differential to generate a neutrophil-to-lymphocyte ratio or systemic immune-inflammation index, serum lactate dehydrogenase and albumin, programmed death-ligand 1 immunohistochemistry, microsatellite stability assessment, and tumor mutational burden. Inclusion of these biomarkers across early-phase immunotherapy clinical trials will capture variation among trials, provide deeper insight into the novel and established therapies, and support improved patient selection and stratification for later-phase clinical trials.

Survival-Inferred Fragility of Statistical Significance in Phase III Oncology Trials

Background

Statistical significance currently defines superiority in phase III oncology trials. However, this practice is increasingly questioned. Here, we estimated the fragility of phase III oncology trials.

Methods

Using Kaplan-Meier curves for the primary endpoints of 230 two-arm superiority phase III oncology trials, we reconstructed data for individual patients. We estimated the survival-inferred fragility index (SIFI) by iteratively flipping the best responder from the experimental arm to the control arm (SIFIB) until the interpretation was changed according to the significance threshold of each trial. Severe fragility was defined by SIFI ≤1%.

Results

This study included 230 trials enrolling 184,752 patients. The median number of patients required to change trial interpretation was 8 (interquartile range, 4 to 19) or 1.4% (interquartile range, 0.7% to 3%) per SIFIB. Estimations of SIFI by multiple methods were largely consistent. For trials with an overall survival primary endpoint, the median SIFIB was 1% (IQR, 0.5% to 1.9%). Severe fragility was found in 87 trials (38%). As a continuous statistic, the original P value-but not its binary significance interpretation-was associated with fragility and severe fragility. Trials with subsequent FDA approval had lower odds of severe fragility. Lastly, the underlying survival model had differential effects on SIFI estimation.

Conclusions

Even among phase III oncology trials, which directly inform patient care, changes in the outcomes of few patients are often sufficient to change statistical significance and trial interpretation. These findings imply that current definitions of statistical significance used in phase III oncology are inadequate to identify replicable findings.

Statistical noise in PD-(L)1 inhibitor trials: unraveling the durable-responder effect

BACKGROUND AND OBJECTIVES

Programmed-death-1/ligand-1 inhibitors (PD-1/L1is) have emerged as pivotal treatments for many cancers. A notable feature of this class of medicines is the dichotomous response pattern: A small (but clinically relevant) percentage of patients (5%-20%) benefit from deep and durable responses resembling functional cures (durable responders), while most patients experience only a modest or negligible response. Accurately predicting durable responders remains elusive due to the lack of a reliable biomarker. Another notable feature of these medicines is that different PD-1/L1 is have obtained statistically significant results, leading to marketing approval for some cancer indications but not for others, with no discernible pattern. These puzzling inconsistencies have generated extensive discussions among oncologists. Proposed (but not entirely convincing) explanations include true underlying differences in efficacy for some types of cancer but not others; or subtle differences in trial design. To investigate a less-explored hypothesis-the durable-responder effect: An initially unidentified group of durable responders generates more statistical noise than anticipated, leading to low-powered randomized controlled trials (RCTs) that report randomly variable results.

STUDY DESIGN

Employing simulation, this investigation divides participants in PD-(L)1i RCTs into two groups: durable responders and patients with a more modest response. Drawing on published data for melanoma, lung and urothelial cancers, multiple prespecified scenarios are replicated 50,000 times, systematically varying the durable-responder percentage from 5% to 20% and the modest-response hazard ratio for overall survival [HR(OS)] from 0.8 to 1.0. This allowed evaluation of the effect of durable responders on power, point estimates of the treatment effect for OS, and the probability of a misleading signal for harm.

RESULTS

When the treatment effect for the modest responders is similar to the comparator arm, statistical power remains below 80%, limiting the ability to reliably detect durable responders. Conversely, there is a material probability of obtaining a statistically significant result that exaggerates the treatment effect by chance. For instance, with an average HR(OS) of 0.93 (corresponding to 5% durable responders), statistically significant trials (7.2%) show an average HR(OS) of 0.77. Additionally, when 5% are durable responders, there is a 20% probability that the HR(OS) will exceed 1.0-suggesting potential harm when none exists.

CONCLUSION

This article adds to the possible explanations for the puzzlingly inconsistent results from PD-(L)1i RCTs. Initially, unidentified durable responders introduce features typical of imprecise, low-powered studies: a propensity for false-negative results; estimates of benefit that might not replicate; and misleading signals for harm.

PLAIN LANGUAGE SUMMARY

Programmed-death-1/ligand-1 (PD-1(L)1) inhibitors are crucial cancer treatments, with global spending expected to surpass $75 billion by 2026. Multiple versions of these medicines are available, all designed to boost the immune system to fight cancer. We would expect them all to work similarly, but clinical trials show mixed results-some seem effective for certain cancers but not others, without a clear pattern. This article uses simulations (virtual trials) to suggest that these inconsistent results may be due to chance, caused by a small group of patients who respond very well to the treatment. Larger trials or specific analysis methods could help reduce the chance effects and provide more robust data for clinician and patient decision-making.

Delayed Separation of Kaplan-Meier Curves is Commonly Observed in Studies of Advanced/Metastatic Solid Tumors Treated with Anti-PD-(L)1 Therapy: Systematic Review and Meta-Analysis

BACKGROUND

Immune checkpoint inhibitor (ICI) Kaplan-Meier (KM) curves often show delayed survival benefit followed by long-term survival in a subgroup of patients. Such outcomes can violate the proportional hazards assumption, leading to a loss of statistical power.

OBJECTIVE

We aimed to determine common trends in delayed separation to inform future ICI clinical trials.

PATIENTS AND METHODS

A literature search was performed using Trialtrove® to identify phase III trials of antiprogrammed cell death (ligand)-1 (anti-PD-[L]1) agents in locally advanced/metastatic solid tumors published between January 2010 and September 2021. The frequency of delayed separation of overall survival (OS) and progression-free survival (PFS) KM curves, correlation between duration of delayed separation in OS/PFS KM curves, and correlation between duration of delayed separation in OS/PFS KM curves with corresponding hazard ratios (HRs) were assessed in all-comer and PD-L1 enriched populations.

RESULTS

Eighty-five studies with OS/PFS KM curves were identified. Most studies showed delayed separation of OS (> 67.9%) and PFS (> 54.5%) KM curves. The correlation between the duration of delayed separation in OS/PFS KM curves was strongest in the PD-L1 enriched population (adjusted R2 = 0.66). No correlation was seen between the duration of delayed separation of OS KM curves and OS HR. A modest correlation was seen between the duration of delayed separation of PFS KM curves and PFS HR in all-comer and PD-L1 enriched populations (adjusted R2 = 0.24 and 0.31, respectively).

CONCLUSIONS

Delayed separation of KM curves was common in clinical trials of anti-PD-(L)1 agents. Understanding delayed separation is key to clinical study designs and assessing outcomes with ICIs.

Detecting the Tumor Prognostic Factors From the YTH Domain Family Through Integrative Pan-Cancer Analysis

Objectives

Emerging evidence suggests that N6-methyladenosine (m6A) methylation plays a critical role in cancers through various mechanisms. This work aims to reveal the essential role of m6A methylation "readers" in regulation of cancer prognosis at the pan-cancer level.

Methods

Herein, we focused on one special protein family of the "readers" of m6A methylation, YT521-B homology (YTH) domain family genes, which were observed to be frequently dysregulated in tumor tissues and closely associated with cancer prognosis. Then, a comprehensive analysis of modulation in cancer prognosis was conducted by integrating RNA sequencing (RNAseq) datasets of YTH family genes and clinical information at the pan-cancer level.

Results

YTH family genes were significantly differentially expressed in most of the cancers, particularly increased in Gastrointestinal cancers, and decreased in Endocrine and Urologic cancers. In addition, they were observed to be associated with overall survival (OS) and disease-specific survival (DSS) with various extent, especially in lower grade glioma (LGG), thyroid cancer (THCA), liver hepatocellular carcinoma (LIHC) and kidney clear cell carcinoma (KIRC), so were some "writers" (METLL3, METLL14, WTAP) and "erasers" (FTO, ALKBH5). Further survival analysis illustrated that YTH family genes specifically YTHScore constructed by combining 5 YTH family genes, as well as RWEScore calculated by combining genes from "readers"-"writers"-"erasers" could dramatically distinguish tumor prognosis in 4 representative cancers. As expected, YTHScore presented an equally comparable prognostic classification with RWEScore. Finally, analysis of immune signatures and clinical characteristics implied that, the activity of the innate immune, diagnostic age, clinical stage, Tumor-Node-Metastasis (TNM) stage and immune types, might play specific roles in modulating tumor prognosis.

Conclusions

The study demonstrated that YTH family genes had the potential to predict tumor prognosis, in which the YTHScore illustrated equal ability to predict tumor prognosis compared to RWEScore, thus providing insights into prognostic biomarkers and therapeutic targets at the pan-cancer level.

Visualizing hypothesis tests in survival analysis under anticipated delayed effects

What can be considered an appropriate statistical method for the primary analysis of a randomized clinical trial (RCT) with a time-to-event endpoint when we anticipate non-proportional hazards owing to a delayed effect? This question has been the subject of much recent debate. The standard approach is a log-rank test and/or a Cox proportional hazards model. Alternative methods have been explored in the statistical literature, such as weighted log-rank tests and tests based on the Restricted Mean Survival Time (RMST). While weighted log-rank tests can achieve high power compared to the standard log-rank test, some choices of weights may lead to type-I error inflation under particular conditions. In addition, they are not linked to a mathematically unambiguous summary measure. Test statistics based on the RMST, on the other hand, allow one to investigate the average difference between two survival curves up to a pre-specified time point τ -a mathematically unambiguous summary measure. However, by emphasizing differences prior to τ , such test statistics may not fully capture the benefit of a new treatment in terms of long-term survival. In this article, we introduce a graphical approach for direct comparison of weighted log-rank tests and tests based on the RMST. This new perspective allows a more informed choice of the analysis method, going beyond power and type I error comparison.

Multi-view contrastive clustering for cancer subtyping using fully and weakly paired multi-omics data

The identification of cancer subtypes is crucial for advancing precision medicine, as it facilitates the development of more effective and personalized treatment and prevention strategies. With the development of high-throughput sequencing technologies, researchers now have access to a wealth of multi-omics data from cancer patients, making computational cancer subtyping increasingly feasible. One of the main challenges in integrating multi-omics data is handling missing data, since not all biomolecules are consistently measured across all samples. Current computational models based on multi-omics data for cancer subtyping often struggle with the challenge of weakly paired omics data. To address this challenge, we propose a novel unsupervised cancer subtyping model named Subtype-MVCC. This model leverages graph convolutional networks to extract and represent low-dimensional features from each omics data type, using intra-view and inter-view contrastive learning approaches. By incorporating a weighted average fusion strategy to unify the dimension of each sample, Subtype-MVCC effectively handles weakly paired multi-omics datasets. Comprehensive evaluations on established benchmark datasets demonstrate that Subtype-MVCC outperforms nine leading models in this domain. Additionally, simulations with varying levels of missing data highlight the model's robust performance in handling weakly paired omics data. The clinical relevance and survival outcomes associated with the identified subtypes further validate the interpretability and reliability of the clustering results produced by Subtype-MVCC.

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

PURPOSE

Survival analyses of novel agents with long-term responders often exhibit differential hazard rates over time. Such proportional hazards violations (PHV) 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 III oncology trials and assess the utility of restricted mean survival time and maximum combination test as additional analyses.

EXPERIMENTAL DESIGN

Clinicaltrials.gov and PubMed were searched to identify two-arm, randomized, phase III 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% confidence interval (CI), 19.7%, 28.5%] comparisons. In multivariable analysis, non-overall survival endpoints [OR, 2.16 (95% CI, 1.21, 3.87); P = 0.009] were associated with higher odds of PHVs, and immunotherapy comparisons [OR 1.94 (95% CI, 0.98, 3.86); P = 0.058] were weakly suggestive of higher odds of PHVs. Few trials with PHVs (25/85, 29.4%) prespecified a statistical plan to account for PHVs. Fourteen trials with PHVs exhibited discordant statistical signals with restricted mean survival time or maximum combination test, of which 10 (71%) reported negative results.

CONCLUSIONS

PHVs are common across therapy types, and attempts to account for PHVs in statistical design are lacking despite the potential for results exhibiting nonproportional hazards to be misinterpreted.

Assurance methods for designing a clinical trial with a delayed treatment effect

An assurance calculation is a Bayesian alternative to a power calculation. One may be performed to aid the planning of a clinical trial, specifically setting the sample size or to support decisions about whether or not to perform a study. Immuno-oncology is a rapidly evolving area in the development of anticancer drugs. A common phenomenon that arises in trials of such drugs is one of delayed treatment effects, that is, there is a delay in the separation of the survival curves. To calculate assurance for a trial in which a delayed treatment effect is likely to be present, uncertainty about key parameters needs to be considered. If uncertainty is not considered, the number of patients recruited may not be enough to ensure we have adequate statistical power to detect a clinically relevant treatment effect and the risk of an unsuccessful trial is increased. We present a new elicitation technique for when a delayed treatment effect is likely and show how to compute assurance using these elicited prior distributions. We provide an example to illustrate how this can be used in practice and develop open-source software to implement our methods. Our methodology has the potential to improve the success rate and efficiency of Phase III trials in immuno-oncology and for other treatments where a delayed treatment effect is expected to occur.

Immune checkpoint inhibitors as subsequent treatment in older adults with non-small cell lung cancer and synchronous brain metastases

Background

Immune checkpoint inhibitors (ICIs) have become the mainstay treatment for non-small cell lung cancer (NSCLC). However, there is a lack of studies assessing ICIs as subsequent treatment in older adults with NSCLC and brain metastasis (BM). This retrospective cohort study compared the real-world survival of older patients with NSCLC and BM at diagnosis [synchronous BM (SBM)] previously treated with chemotherapy receiving ICI versus chemotherapy as subsequent treatment.

Methods

Patients with NSCLC and SBM ≥65 years previously treated with chemotherapy were identified using the SEER-Medicare database (2010-2019). Patients receiving new chemotherapy and/or Food and Drug Administration (FDA)-approved ICIs as second/third-line treatment were included, excluding those ever-receiving targeted therapies. Each ICI patient was matched to one chemotherapy patient by time to subsequent treatment (within ±30 days) from diagnosis. Overall survival (OS) time was measured from the start of subsequent treatment to death, censored at disenrollment from Medicare Part A/B, enrollment in Part C, or end of study (December 31, 2019), whichever came first. OS curves were estimated and compared using the Kaplan-Meier (KM) method and log-rank test. Hazard ratio (HR) was estimated using a multivariable-adjusted Cox proportional hazards model.

Results

Matched cohorts included 546 patients [273 in each group; median age 71 (range, 65-87) years]. ICI patients were older, more likely non-Hispanic, with squamous cell carcinoma, and liver metastasis compared to chemotherapy. KM estimated better survival in ICI than chemotherapy {median survival: 209 days [95% confidence interval (CI): 160-275] vs. 155 days (95% CI: 135-187); log-rank P<0.001}. ICI was associated with a lower adjusted hazard of death [HR =0.63; 95% CI: 0.52-0.75; P<0.001] compared to subsequent chemotherapy treatment.

Conclusions

In this population-based study of older patients with NSCLC and SBM previously treated with chemotherapy, subsequent treatment with ICI was associated with improved survival compared to chemotherapy.

Methods for non-proportional hazards in clinical trials: A systematic review

For the analysis of time-to-event data, frequently used methods such as the log-rank test or the Cox proportional hazards model are based on the proportional hazards assumption, which is often debatable. Although a wide range of parametric and non-parametric methods for non-proportional hazards has been proposed, there is no consensus on the best approaches. To close this gap, we conducted a systematic literature search to identify statistical methods and software appropriate under non-proportional hazard. Our literature search identified 907 abstracts, out of which we included 211 articles, mostly methodological ones. Review articles and applications were less frequently identified. The articles discuss effect measures, effect estimation and regression approaches, hypothesis tests, and sample size calculation approaches, which are often tailored to specific non-proportional hazard situations. Using a unified notation, we provide an overview of methods available. Furthermore, we derive some guidance from the identified articles.

Sample size calculation for mixture model based on geometric average hazard ratio and its applications to nonproportional hazard

With the advent of cancer immunotherapy, some special features including delayed treatment effect, cure rate, diminishing treatment effect and crossing survival are often observed in survival analysis. They violate the proportional hazard model assumption and pose a unique challenge for the conventional trial design and analysis strategies. Many methods like cure rate model have been developed based on mixture model to incorporate some of these features. In this work, we extend the mixture model to deal with multiple non-proportional patterns and develop its geometric average hazard ratio (gAHR) to quantify the treatment effect. We further derive a sample size and power formula based on the non-centrality parameter of the log-rank test and conduct a thorough analysis of the impact of each parameter on performance. Simulation studies showed a clear advantage of our new method over the proportional hazard based calculation across different non-proportional hazard scenarios. Moreover, the mixture modeling of two real trials demonstrates how to use the prior information on the survival distribution among patients with different biomarker and early efficacy results in practice. By comparison with a simulation-based design, the new method provided a more efficient way to compute the power and sample size with high accuracy of estimation. Overall, both theoretical derivation and empirical studies demonstrate the promise of the proposed method in powering future innovative trial designs.

Prediction and verification of the prognostic biomarker SLC2A2 and its association with immune infiltration in gastric cancer

Gastric cancer (GC) is the fifth most common cause of cancer-associated deaths; however, its treatment options are limited. Despite clinical improvements, chemotherapy resistance and metastasis are major challenges in improving the prognosis and quality of life of patients with GC. Therefore, effective prognostic biomarkers and targets associated with immunological interventions need to be identified. Solute carrier family 2 member 2 (SLC2A2) may serve a role in tumor development and invasion. The present study aimed to evaluate SLC2A2 as a prospective prognostic marker and chemotherapeutic target for GC. SLC2A2 expression in several types of cancer and GC was analyzed using online databases, and the effects of SLC2A2 expression on survival prognosis in GC were investigated. Clinicopathological parameters were examined to explore the association between SLC2A2 expression and overall survival (OS). Associations between SLC2A2 expression and immune infiltration, immune checkpoints and IC50 were estimated using quantification of the tumor immune contexture from human RNA-seq data, the Tumor Immune Estimation Resource 2.0 database and the Genomics of Drug Sensitivity in Cancer database. Differential SLC2A2 expression and the predictive value were validated using the Human Protein Atlas, Gene Expression Omnibus, immunohistochemistry and reverse transcription-quantitative PCR. SLC2A2 expression was downregulated in most types of tumor but upregulated in GC. Functional enrichment analysis revealed an association between SLC2A2 expression and lipid metabolism and the tumor immune microenvironment. According to Gene Ontology term functional enrichment analysis, SLC2A2-related differentially expressed genes were enriched predominantly in 'chylomicron assembly', 'plasma lipoprotein particle assembly', 'high-density lipoprotein particle', 'chylomicron', 'triglyceride-rich plasma lipoprotein particle', 'very-low-density lipoprotein particle'. 'intermembrane lipid transfer activity', 'lipoprotein particle receptor binding', 'cholesterol transporter activity' and 'intermembrane cholesterol transfer activity'. In addition, 'cholesterol metabolism', and 'fat digestion and absorption' were significantly enriched in the Kyoto Encyclopedia of Genes and Genomes pathway analysis. Patients with GC with high SLC2A2 expression had higher levels of neutrophil and M2 macrophage infiltration and a significant inverse correlation was observed between SLC2A2 expression and MYC targets, tumor mutation burden, microsatellite instability and immune checkpoints. Furthermore, patients with high SLC2A2 expression had worse prognosis, including OS, disease-specific survival and progression-free interval. Multivariate regression analysis demonstrated that SLC2A2 could independently prognosticate GC and the nomogram model showed favorable performance for survival prediction. SLC2A2 may be a prospective prognostic marker for GC. The prediction model may improve the prognosis of patients with GC in clinical practice, and SLC2A2 may serve as a novel therapeutic target to provide immunotherapy plans for GC.

The long and winding road: Adjuvant therapy for early-stage hepatocellular carcinoma

Despite being the first positive phase 3 trial in the adjuvant setting for early-stage hepatocellular carcinoma, the IMbrave050 study raises a number of questions regarding patient selection, endpoint robustness, and the balance between efficacy and acceptable toxicity.1.

The Inverse Log-Rank Test: A Versatile Procedure for Late Separating Survival Curves

Often in the planning phase of a clinical trial, a researcher will need to choose between a standard versus weighted log-rank test (LRT) for investigating right-censored survival data. While a standard LRT is optimal for analyzing evenly distributed but distinct survival events (proportional hazards), an appropriately weighted LRT test may be better suited for handling non-proportional, delayed treatment effects. The "a priori" misspecification of this alternative may result in a substantial loss of power when determining the effectiveness of an experimental drug. In this paper, the standard unweighted and inverse log-rank tests (iLRTs) are compared with the multiple weight, default Max-Combo procedure for analyzing differential late survival outcomes. Unlike combination LRTs that depend on the arbitrary selection of weights, the iLRT by definition is a single weight test and does not require implicit multiplicity correction. Empirically, both weighted methods have reasonable flexibility for assessing continuous survival curve differences from the onset of a study. However, the iLRT may be preferable for accommodating delayed separating survival curves, especially when one arm finishes first. Using standard large-sample methods, the power and sample size for the iLRT are easily estimated without resorting to complex and timely simulations.

Two-sample inference procedures under nonproportional hazards

We introduce a new two-sample inference procedure to assess the relative performance of two groups over time. Our model-free method does not assume proportional hazards, making it suitable for scenarios where nonproportional hazards may exist. Our procedure includes a diagnostic tau plot to identify changes in hazard timing and a formal inference procedure. The tau-based measures we develop are clinically meaningful and provide interpretable estimands to summarize the treatment effect over time. Our proposed statistic is a U-statistic and exhibits a martingale structure, allowing us to construct confidence intervals and perform hypothesis testing. Our approach is robust with respect to the censoring distribution. We also demonstrate how our method can be applied for sensitivity analysis in scenarios with missing tail information due to insufficient follow-up. Without censoring, Kendall's tau estimator we propose reduces to the Wilcoxon-Mann-Whitney statistic. We evaluate our method using simulations to compare its performance with the restricted mean survival time and log-rank statistics. We also apply our approach to data from several published oncology clinical trials where nonproportional hazards may exist.

Real-World Survival of First-Line Immune Checkpoint Inhibitor Treatment Versus Chemotherapy in Older Patients With Non-Small-Cell Lung Cancer and Synchronous Brain Metastases

PURPOSE

This study assessed real-world survival among older patients with non-small-cell lung cancer (NSCLC) and brain metastases (BMs) at diagnosis (synchronous BM [SBM]) receiving first-line immune checkpoint inhibitors (ICIs) compared with chemotherapy only.

METHODS

Patients with NSCLC and SBM age 65 years or older at diagnosis from 2010 to 2019 SEER-Medicare database and received US Food and Drug Administration-approved ICIs (pembrolizumab/nivolumab/ipilimumab/atezolizumab/durvalumab/cemiplimab) and/or chemotherapy (platinum-based doublets/taxane/pemetrexed/gemcitabine) as first-line systemic treatment were included, excluding those with no cranial radiation or ever being treated with targeted therapies. Overall survival time was from the start of systemic treatment (ICI/chemotherapy) to death, censored at disenrollment from Medicare part A/B, enrollment in part C, or end of the study period (December 31, 2019). Kaplan-Meier (KM) survival curves were compared between treatment groups using the log-rank test. Multivariable Cox proportional hazards (CPH) model was used to estimate hazard ratio (HR) between groups, adjusting for patients' sociodemographic and clinical characteristics.

RESULTS

The study included 1,481 patients (1,303 chemotherapy and 178 ICI). The median (range) age was 71 (65-91) years. First-line ICI patients were more likely to be older, live in urban areas, and less likely to be non-White than the chemotherapy group. KM estimates showed that survival curves initially overlapped but diverged approximately 6 months after initiating first-line systemic treatment (median survival [95% CI]: ICI, 190 [131 to 303] days versus chemotherapy, 189 [177 to 201] days), with ICI showing a better survival than the chemotherapy group (log-rank test P < .0001). First-line ICI was associated with a lower risk of death compared with chemotherapy in adjusted CPH model (HR [95% CI], 0.67 [0.55 to 0.80]; P < .0001).

CONCLUSION

Among older patients with NSCLC and SBM, first-line ICI use was associated with improved survival occurring 6 months after treatment initiation compared with chemotherapy only.

Challenges, Complexities, and Considerations in the Design and Interpretation of Late-Phase Oncology Trials

Optimal management of cancer patients relies heavily on late-phase oncology randomized controlled trials. A comprehensive understanding of the key considerations in designing and interpreting late-phase trials is crucial for improving subsequent trial design, execution, and clinical decision-making. In this review, we explore important aspects of late-phase oncology trial design. We begin by examining the selection of primary endpoints, including the advantages and disadvantages of using surrogate endpoints. We address the challenges involved in assessing tumor progression and discuss strategies to mitigate bias. We define informative censoring bias and its impact on trial results, including illustrative examples of scenarios that may lead to informative censoring. We highlight the traditional roles of the log-rank test and hazard ratio in survival analyses, along with their limitations in the presence of nonproportional hazards as well as an introduction to alternative survival estimands, such as restricted mean survival time or MaxCombo. We emphasize the distinctions between the design and interpretation of superiority and noninferiority trials, and compare Bayesian and frequentist statistical approaches. Finally, we discuss appropriate utilization of phase II and phase III trial results in shaping clinical management recommendations and evaluate the inherent risks and benefits associated with relying on phase II data for treatment decisions.

Efficacy and safety of frontline systemic therapy for advanced HCC: A network meta-analysis of landmark phase III trials

Background & Aims

Direct comparisons across first-line regimens for advanced hepatocellular carcinoma are not available. We performed a network metanalysis of phase III of trials to compare first-line systemic treatments for hepatocellular carcinoma in terms of overall survival (OS), progression-free survival (PFS), objective response rate, disease control rate, and incidence of adverse events (AEs).

Methods

After performing a literature review from January 2008 to September 2022, we screened 6,329 studies and reviewed 3,009 studies, leading to identification of 15 phase III trials for analysis. We extracted odds ratios for objective response rate and disease control rate, relative risks for AEs, and hazard ratios (HRs) with 95% CIs for OS and PFS, and used a frequentist network metanalysis, with fixed-effect multivariable meta-regression models to estimate the indirect pooled HRs, odds ratios, relative risks, and corresponding 95% CIs, considering sorafenib as reference.

Results

Of 10,820 included patients, 10,444 received active treatment and 376 placebo. Sintilimab + IBI350, camrelizumab + rivoceranib, and atezolizumab + bevacizumab provided the greatest reduction in the risk of death compared with sorafenib, with HRs of 0.57 (95% CI 0.43-0.75), 0.62 (95% CI 0.49-0.79), and 0.66 (95% CI 0.52-0.84), respectively. Considering PFS, camrelizumab + rivoceranib and pembrolizumab + lenvatinib were associated with the greatest reduction in the risk of PFS events compared with sorafenib, with HRs of 0.52 (95% CI 0.41-0.65) and 0.52 (95% CI 0.35-0.77), respectively. Immune checkpoint inhibitor (ICI) monotherapies carried the lowest risk for all-grade and grade ≥3 AEs.

Conclusions

The combinations of ICI + anti-vascular endothelial growth factor, and double ICIs lead to the greatest OS benefit compared with sorafenib, whereas ICI + kinase inhibitor regimens are associated with greater PFS benefit at the cost of higher toxicity rates.

Impact and Implications

In the last few years, many different therapies have been studied for patients with primary liver cancer that cannot be treated with surgery. In these cases, anticancer drugs (alone or in combination) are given with the intent to keep the cancer at bay and, ultimately, to prolong survival. Among all the therapies that have been investigated, the combination of immunotherapy (drugs that boost the immune system against the cancer) and anti-angiogenic agents (drugs that act on tumoural vessels) has appeared the best to improve survival. Similarly, the combination of two types of immunotherapies that activate the immune system at different levels has also shown positive results.

Systematic Review Registration

PROSPERO CRD42022366330.

Subtype-DCC: decoupled contrastive clustering method for cancer subtype identification based on multi-omics data

Due to the high heterogeneity and complexity of cancers, patients with different cancer subtypes often have distinct groups of genomic and clinical characteristics. Therefore, the discovery and identification of cancer subtypes are crucial to cancer diagnosis, prognosis and treatment. Recent technological advances have accelerated the increasing availability of multi-omics data for cancer subtyping. To take advantage of the complementary information from multi-omics data, it is necessary to develop computational models that can represent and integrate different layers of data into a single framework. Here, we propose a decoupled contrastive clustering method (Subtype-DCC) based on multi-omics data integration for clustering to identify cancer subtypes. The idea of contrastive learning is introduced into deep clustering based on deep neural networks to learn clustering-friendly representations. Experimental results demonstrate the superior performance of the proposed Subtype-DCC model in identifying cancer subtypes over the currently available state-of-the-art clustering methods. The strength of Subtype-DCC is also supported by the survival and clinical analysis.