Adjuvant Nivolumab After Cystectomy in a PD-L1-Negative Tumor
CheckMate 274 supports adjuvant nivolumab broadly, but its clearest, most consistent benefit sat in the PD-L1-positive subgroup — his tumor is PD-L1-negative, and the FDA's all-comers approval doesn't settle whether the benefit actually reaches a patient like him.
P.A., a 66-year-old woman, retired from three decades as a hospital pharmacist eight months ago, a career she says taught her to read a drug's actual trial data before trusting a headline indication — a habit that has made this exact conversation more pointed than it might be with another patient. Her radical cystectomy for pT3N1 high-grade urothelial carcinoma was six weeks ago, margins clear, no residual disease on postoperative imaging. Her disease meets CheckMate 274's high-risk enrollment criteria outright by stage and nodal involvement, and tumor testing came back PD-L1 combined positive score under 1% — negative by the trial's own biomarker cutoff.
CheckMate 274 showed a disease-free survival benefit for adjuvant nivolumab across its overall population, which is what earned its broad FDA approval, but the effect was substantially larger and more consistent in the PD-L1-positive subgroup specifically; the European Medicines Agency, reviewing the same data, restricted its own approval to PD-L1-positive tumors only, judging the negative-subgroup signal too weak to support routine use. She has read both agencies' reasoning herself and has arrived at her appointment with a specific question: whether the FDA's broader label reflects a real benefit her tumor is likely to share, or whether it reflects a regulatory choice not to subdivide an already-positive overall trial result. CheckMate 274's own published subgroup analysis reported a disease-free survival hazard ratio meaningfully more favorable in the PD-L1-positive group than in the PD-L1-negative group, though the negative subgroup's own confidence interval still trended toward benefit without reaching the same statistical clarity — a pattern she described, having read the forest plot herself before this visit, as "exactly wide enough to argue about." She has asked specifically to see that subgroup breakdown itself rather than a summary of it, along with the actual text of the EMA's restricted indication, before deciding anything — a request her surgeon has honored by pulling both documents ahead of today's visit rather than paraphrasing them from memory.
Six weeks after cystectomy, reviewing the biomarker result
She meets CheckMate 274's high-risk criteria by stage and nodal status, and the trial's overall population showed a real disease-free survival benefit for adjuvant nivolumab — that's what the FDA approval is built on, and it's an all-comers approval, not restricted by PD-L1 status. I'd offer it on that basis.
I'd want to be direct with her about something the FDA label doesn't say plainly: the EMA reviewed the identical trial data and restricted their own approval to PD-L1-positive tumors specifically, because they judged the negative-subgroup benefit too weak to support routine use. Her tumor is PD-L1-negative. That's not a technicality — it's two major regulators looking at the same numbers and disagreeing about exactly her situation.
An all-comers approval tells us the overall trial met its endpoint. It doesn't settle whether every biomarker subgroup inside that overall result benefited equally — and here we actually have a second regulator's independent read saying it probably didn't.
She's told us directly she wants to see the actual data, not just be told which label applies here. I think that's exactly right for a decision like this one — I'd walk her through both agencies' reasoning and the subgroup numbers themselves, and let her weigh the real uncertainty against nivolumab's own toxicity risk herself, rather than either of us deciding it for her.
Agreed: both regulatory reviews and the underlying subgroup data would be presented to her directly at her next visit, with her own decision determining whether nivolumab starts.
Not agreed between the two oncologists on what they would recommend if asked directly:
Would recommend starting nivolumab, treating the FDA's all-comers approval as the operative standard of care.
Would recommend surveillance alone, treating the EMA's biomarker-restricted approval as the more defensible reading of the same data.