Use of biomarkers and surrogate endpoints in health technology assessment (HTA): efficient decision-making for allocation of pharmacological treatment strategies in subpopulations of metastatic colorectal cancer patients
When evaluating new cancer therapies in clinical trials, it may take significant time toestimate their effectiveness on overall survival (OS), which is typically an outcome of central interest to regulatory decision-makers. To expedite access to new therapies for patients, regulatory agencies often make their decisions based on treatment effectiveness measured on surrogate outcomes, such as progression free survival (PFS) or tumour response (TR). For such decisions to be robust, a surrogate endpoint needs to be a valid predictor of OS. The aim of the thesis was to investigate the impact of a treatment’s mechanism of action and patients’ Kirsten rat sarcoma (KRAS) status on surrogacy patterns in metastatic colorectal cancer (mCRC). Putative surrogate endpoints considered here were PFS as a surrogate endpoint for OS, and TR as a surrogate for PFS and OS. The surrogate relationships were evaluated across all RCTs, by KRAS status, treatment class, and treatment contrast; using a range of bivariate meta-analytic methods for surrogate endpoint evaluation. A case study in mCRC was used to explore the potential benefits that borrowing of information across trials and treatment classes can provide to inform decision-making.
PFS appeared to be a good surrogate endpoint for OS. The surrogate relationship for TR with PFS and OS varied from exhibiting moderate evidence to no evidence to support a surrogate relationship depending on the data set. There was some evidence of the impact of the mechanism of action on the strength of the surrogacy patterns in mCRC, but little evidence of the impact of KRAS status on the validity of surrogate endpoints. The use of these methods within a health technology assessment setting can be beneficial for confirming the validity of a surrogate endpoint, making predictions for OS, and ultimately allowing a trial with premature or no OS data to be assessed by decision-making bodies.
Supervisor(s)Sylwia Bujkiewicz; Michael Sweeting; Sam Khan
Date of award2023-12-18
Author affiliationHealth Sciences
Awarding institutionUniversity of Leicester