Time to deterioration of patient-reported outcomes as a surrogate of overall survival: a meta-analysis

Prognostic value of baseline EORTC QLQ-C30 scores for overall survival across 46 clinical trials covering 17 cancer types: a validation study

Background

A pooled data analysis by Quinten et al. (2009) found three European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) health-related quality of life (HRQoL) scales to be prognostic for survival: physical functioning, pain and appetite loss. This study aims to replicate these findings in an independent data set comprising a broader cancer population.

Methods

Data were obtained from 46 clinical trials across three cancer research networks conducted between 1996 and 2013 that assessed HRQoL using the EORTC QLQ-C30. A stratified Cox proportional hazards model was employed to assess the prognostic significance of baseline QLQ-C30 scale scores on overall survival, adjusting for socio-demographic and clinical variables. Stepwise model selection was done at 5% significance level. Model stability and prognostic accuracy were evaluated via bootstrapping and the C index respectively.

Findings

Data from 16,210 patients reporting HRQoL at baseline, spanning 17 cancer types, was used. The stratified multivariable model confirmed that better physical functioning (hazard ratio [HR], 0.94; 95% confidence interval [CI], 0.93-0.96), lower pain (HR, 1.02; 95% CI, 1.01-1.03), and appetite loss (HR, 1.04; 95% CI, 1.03-1.05) were significantly associated with survival. Additionally, global health status/QoL, dyspnoea, emotional and cognitive functioning were found to be prognostic for survival. This final model, encompassing sociodemographic, clinical, and HRQoL variables, achieved a corrected C index of 0.74, marking a 48% enhancement in discriminatory ability. Bootstrap evaluation indicated no major instability issues.

Interpretation

These results support previous findings that baseline physical functioning, pain, and appetite loss scores, along with four other scales from the EORTC QLQ-C30, predict survival in cancer patients.

Funding

EORTC Quality of Life Group.

Patient-reported outcome domains in multiple myeloma randomized controlled trials and association with survival outcomes

Patient-reported outcomes (PROs) are crucial endpoints in multiple myeloma (MM) randomized controlled trials (RCTs), yet there is significant variability in their methodology and reporting. Our study aimed to (a) identify the most commonly pre-specified PRO domains in MM RCTs and those most responsive to modern therapies, and (b) examine the association between PROs and progression-free survival (PFS)/overall survival (OS). We performed a systematic review of MM RCTs that used EORTC QLQ-C30 and published between 01/2014-06/2023. The association between PFS/OS and PRO was explored using Fisher's exact test or Pearson's Chi-squared test. Thirty-five RCTs were identified, with PROs as secondary or exploratory endpoints in all studies. About one-third of RCTs (n=11, 31.4%) pre-specified at least one EORTC QLQ-C30 domain, with the most common domains being Global health status/Quality of life (GHS/QoL) (n = 10, 90.9%), Physical Functioning (n = 6, 54.5%), Fatigue (n = 6, 54.5%), and Pain (n = 6, 54.5%). A statistically significant and/or clinically meaningful difference in at least one EORTC QLQ-C30 domain between arms was seen in 23/35 trials (65.7%), with the most common domains showing improvement being GHS/QoL (12/23 trials), Pain (11/23 trials), Fatigue (9/23 trials), and Physical Functioning (9/23 trials). PRO was noted to be concordant with PFS in 19/33 (57.6%) trials (p = 0.398), and with OS in 22/31 (71%) trials (p = 0.018). Our study identified key PRO domains that can be potentially used as primary endpoint in MM RCTs. Additionally, significant association between PROs and OS highlight the importance of integrating PROs to better capture treatment efficacy.

Objective response rate and progression-free survival as surrogates for overall survival treatment effect: A meta-analysis across diverse tumour groups and contemporary therapies

BACKGROUND

Overall survival (OS) results from randomized control trials (RCT) provide the strongest evidence for efficacy of anti-cancer treatments but can take a considerable amount of time to mature. Progression free survival (PFS) and objective response rate (ORR) are used as an early surrogate of OS treatment effect however their validity remains unclear. Our study aims to comprehensively evaluate ORR and PFS as surrogates for OS treatment effect across tumor groups and treatment types.

MATERIAL AND METHODS

Phase 3 RCTs in solid malignancies that reported OS/PFS and ORR published between 1st of January 2010 and 30th of June 2022 were evaluated. The relationship of surrogate endpoints and OS treatment effect was assessed via weighted linear regression. The coefficient of determination (R2) quantified the fit of the regression model.

RESULTS

675 phase 3 RCT comprising of 350 112 patients were analysed. ORR (R2 of 0.10) and PFS (R2 of 0.38) were poor surrogate markers of OS treatment effect. The strength of surrogacy differed within treatment and tumour groups. PFS had the highest R2 for chemotherapy (0.56) and lowest for targeted therapy (0.40). PFS had the highest level of surrogacy for melanoma (R2 = 0.72) and pancreatic cancer (R2 = 0.70) compared to other tumour groups. Importantly ORR and PFS were also poorly correlated to each other (R2 = 0.33).

CONCLUSIONS

ORR and PFS were poor trial-level surrogate markers of OS. The surrogacy performance of ORR and PFS differed by treatment and malignancy sub-type.