How serious is bias in effect estimation in randomised trials with survival data given risk heterogeneity and informative censoring?

How hazardous are hazard ratios? An empirical investigation of individual patient data from 27 large randomized clinical trials

The use of hazard ratios as the standard treatment effect estimators for randomized trials with time-to-event outcomes has been the subject of repeated criticisms in recent years, e.g., for its non-collapsibility or with respect to (causal) interpretation. Another important issue is the built-in selection bias, which arises when the treatment is effective and when there are unobserved or not included prognostic factors that influence time-to-event. In these cases, the hazard ratio has even been termed "hazardous" because it is estimated from groups that increasingly differ in their (unobserved or omitted) baseline characteristics, yielding biased treatment estimates. We therefore adapt the Landmarking approach to assess the effect of ignoring a gradually increasing proportion of early events on the estimated hazard ratio. We propose an extension called "Dynamic Landmarking". This approach is based on successive deletion of observations, refitting Cox models and balance checking of omitted but observed prognostic factors, to obtain a visualization that can indicate built-in selection bias. In a small proof-of-concept simulation, we show that our approach is valid under the given assumptions. We further use "Dynamic Landmarking" to assess the suspected selection bias in the individual patient data sets of 27 large randomized clinical trials (RCTs). Surprisingly, we find no empirical evidence of selection bias in these RCTs and thus conclude that the supposed bias of the hazard ratio is of little practical relevance in most cases. This is mainly due to treatment effects in RCTs being small and the patient populations being homogeneous, e.g., due to inclusion and exclusion criteria.

Comprehensive Analysis of TP53 and KEAP1 Mutations and Their Impact on Survival in Localized- and Advanced-Stage NSCLC

INTRODUCTION

TP53 and KEAP1 are frequently mutated in NSCLC, but their prognostic value is ambiguous, particularly in localized stage tumors.

METHODS

This retrospective cohort study included a total of 6297 patients with NSCLC who were diagnosed between November 1998 and February 2020. The primary end point was overall survival. Patients were diagnosed in a central pathology laboratory as part of the Network Genomic Medicine collaboration, encompassing more than 300 lung cancer-treating oncology centers in Germany. All patients underwent molecular testing, including targeted next-generation panel sequencing and in situ hybridization.

RESULTS

A total of 6297 patients with NSCLC were analyzed. In 1518 surgically treated patients (Union for International Cancer Control [UICC] I-IIIA), truncating TP53 mutations and KEAP1 mutations were independent negative prognostic markers in multivariable analysis (hazard ratio [HR]_{TP53truncating} = 1.43, 95% confidence interval [CI]: 1.07-1.91, p = 0.015; HR_KEAP1mut = 1.68, 95% CI:1.24-2.26, p = 0.001). Consistently, these mutations were associated with shorter disease-free survival. In 4779 patients with advanced-stage (UICC IIIB-IV) tumors, TP53 mutations did not predict outcome in univariable analysis. In contrast, KEAP1 mutations remained a negative prognostic factor (HR_KEAP1mut = 1.40, 95% CI: 1.23-1.61, p < 0.001) in patients with advanced-stage tumors. Furthermore, those with KEAP1-mutant tumors with co-occurring TP53 missense mutations had longer overall survival than those with KEAP1-mutant tumors with wild-type or truncating TP53 mutations.

CONCLUSIONS

This study found that TP53 and KEAP1 mutations were prognostic for localized and advanced-stage NSCLC. The increased relative hazard of harboring TP53 or KEAP1 mutations was comparable to an increase in one UICC stage. Our data suggest that molecular stratification on the basis of TP53 and KEAP1 mutation status should be implemented for localized and advanced-stage NSCLC to optimize and modify clinical decision-making.