
OS Benefit, Patient Communication, and Frontline Risk in DLBCL
Dr. Graff addresses whether the absence of overall survival benefit from either POLARIX or frontMIND should discourage adoption.
Dr. Graff addresses whether the absence of overall survival benefit from either POLARIX or frontMIND should discourage adoption. She argues no: both trials are young, and expecting an overall survival signal at early readouts is unrealistic given the heterogeneous DLBCL patient population, the multiple subsequent therapies now available at relapse including CAR-T, and the fact that no frontline randomized large B-cell lymphoma trial has ever been powered with overall survival as its primary endpoint.
She draws an analogy to the ECHELON-1 experience in Hodgkin lymphoma, where brentuximab vedotin plus adriamycin, vinblastine, and dacarbazine showed modified progression-free survival advantage without immediate overall survival benefit, which emerged at 6-year follow-up. She argues PFS improvements in potentially curable disease will eventually translate, and that avoiding relapse itself, with all its associated toxicity, disruption, and cost, represents a meaningful win for patients even absent overall survival data.
Dr. Lunning emphasizes that the vast majority of patients relapsing after Pola-R-CHP or Tafa-Len-R-CHOP will be early relapsers or primary refractory, most of whom will be directed toward bispecific antibodies or anti-CD19 CAR-T cell therapy rather than platinum-based salvage and autologous transplant. He anticipates that the concept of "need for additional lymphoma therapy" as a clinical endpoint will gain importance in frontMIND, as it did in POLARIX.
When counseling patients who ask whether a new regimen will help them live longer, Dr. Westin frames the conversation around preventing the need for disruptive second-line therapies including CAR-T, noting that even when no overall survival benefit exists, avoiding relapse treatment is a clinically meaningful goal that improves quality of life.
Related to this article








