Multi-state models for the analysis of survival data in radiation therapy.

Multi-state models for the analysis of survival data in radiation therapy.

Most of the clinical trials with survival data involving radiation therapy are analysed using Cox regression model. Each features of the survival process are analysed separately (e.g. progression free survival, cancer specific survival, overall survival…). In this setting, multi-state models (MSM) are an asset when overall survival is highly influenced by an intermediate event of cancer progression, giving insight on the interplay between the different states ( e.g. “free from recurrence”, “local progression”, ”quality of life deterioration” or “death”). This study aims to illustrate trough a well-known example, what kind of information a MSM is able to provide, compare to the usual regression models. In the setting of the AVAglio trial, MSM should give a more detailed analysis of the impact of progression on overall survival according to the treatment arm. Such data could come up with insight to guide future clinical trials in radiotherapy by encouraging the use of extended statistical methods such as MSM for further data exploitation.

[{ "PostingID": 19661, "Title": "ROCHE-BO21990", "Description": "A randomized, double-blind, placebocontrolled, multicenter Phase III trial of bevacizumab (Bv), temozolomide (TMZ) and radiotherapy (RT), followed by Bv and TMZ versus placebo (Pl), TMZ and RT followed by Pl and TMZ in patients with newly diagnosed glioblastoma." }]