A narrative review of survival normalization in non-Hodgkin lymphoma: useful for better patient counselling and an endpoint for clinical trials?

Abstract: As numerous new therapies against non-Hodgkin lymphoma (NHL) are under investigation, there is an increasing interest in surrogate endpoints (SEPs) that can reduce the costs and duration of clinical trials and thereby accelerate delivery of new treatments to patients with unmet medical needs. The time to normalization of overall survival (OS) is an endpoint that is increasingly explored for this purpose. The patient survival is considered normalized when it matches the corresponding survival of a matched background population. The time point where this occurs is also useful for patient counselling and for developing rationale clinical follow-up programs. Survival normalization has been investigated in both aggressive and indolent NHLs with some results indicating survival normalization after 24 months of event-free survival (EFS24) in diffuse large B-cell lymphoma (DLBCL) patients and for patients with follicular lymphoma (FL) with complete response after 30 months (CR30). Here, we review the concept of survival normalization endpoints and their potential as SEPs in future NHL trials. As intervention effects on SEPs should adequately predict effects on conventional endpoints, validation generally requires extensive analysis of data from multiple trials, and thus only few validation studies exist within cancer. The validity of EFS24 and CR30 as surrogates for OS and progression-free survival (PFS), respectively, was investigated in a series of trials from the Follicular Lymphoma Analysis of Surrogate Hypothesis (FLASH) and Surrogate Endpoints for Aggressive Lymphoma (SEAL) consortiums. The results suggested a strong correlation between CR30 and PFS results in FL, whereas the correlation between intervention effects on the EFS24 endpoint and OS in DLBCL did not met prespecified thresholds despite clear correlation. In conclusion, survival normalization is a clinically important concept, but more research is needed before the EFS24 endpoint can be applied in future DLBCL trials. On the other hand, results suggest that CR30 may be a suitable SEP for PFS in future FL trials.