Adjuvant Immunotherapy After Nephrectomy, Before the Survival Data Matured
A patient possibly already cured by surgery alone is being offered a year of adjuvant immunotherapy on the strength of a disease-free survival benefit whose overall survival data hadn't yet matured when the decision had to be made.
L.S., a 55-year-old woman, coaches her son's competitive swim team on weekday evenings, a commitment she's kept without missing a practice through her entire diagnosis and recovery. Her right partial nephrectomy for a 9cm clear cell renal cell carcinoma with sarcomatoid features and positive regional lymph nodes was six weeks ago; margins were clear, and postoperative imaging shows no evidence of residual or metastatic disease. Her tumor's size, nodal involvement, and sarcomatoid differentiation together place her in the high-risk category KEYNOTE-564 enrolled, the trial that established a disease-free survival benefit for one year of adjuvant pembrolizumab after resection in exactly this population.
This case is set at the decision point it was, in 2023, before KEYNOTE-564's overall survival data matured: the disease-free survival benefit was real and statistically significant, but whether it would translate into a survival advantage, rather than simply delaying detection of recurrences that might have been treated just as effectively at relapse, was genuinely unknown at the time. It has since been answered: the third interim analysis, published in 2024, reported a significant overall survival benefit — hazard ratio 0.62, 48-month survival 91.2% against 86.0% — the first for any adjuvant regimen in renal cell carcinoma. The debate below is preserved as it stood, because the reasoning under uncertainty is the point; the outcome should not be read as still open. Against that uncertain benefit sits a real and partly irreversible cost: pembrolizumab's immune-related adverse events, most commonly thyroid dysfunction and colitis but occasionally adrenal insufficiency or pneumonitis, can persist long after the drug is stopped, and some fraction of resected patients in her risk category are, by definition, already cured by surgery alone and would be exposed to that risk for no benefit at all — a fraction the trial's own design cannot identify in advance for any individual patient. In KEYNOTE-564, immune-mediated events of any grade occurred in 36.5% of patients against 7.3% on placebo, though grade 3 or 4 immune-mediated events stayed under 10%. The endocrinopathies are the ones that matter to a woman weighing a year of therapy against a chance she is already cured, since thyroid and adrenal dysfunction can require hormone replacement indefinitely after the drug is stopped — a low-probability, long-duration harm rather than a high-probability, temporary one. She raised her son's upcoming competitive swim season herself at this visit, noting that a colitis flare or a fatigue-inducing endocrinopathy during his qualifying meets was, to her, a very different kind of cost than an abstract toxicity percentage on a consent form.
Six weeks after surgery, before the adjuvant window closes
I'd recommend starting pembrolizumab. KEYNOTE-564 showed a real, statistically significant disease-free survival benefit in exactly her risk category — sarcomatoid features, nodal involvement, large tumor size. We've adopted disease-free-survival-driven adjuvant therapies before survival data matured in other settings, and waiting risks missing the window where the drug is actually effective.
The overall survival data aren't in yet — that third interim analysis is still years away as we sit here — and that matters specifically because some real fraction of patients in her risk category are already cured by surgery alone. We can't tell which ones. A year of pembrolizumab carries a real risk of irreversible immune-related toxicity — permanent thyroid or adrenal dysfunction, occasionally worse — and for an already-cured patient, that's pure harm with no benefit at all.
A disease-free survival benefit tells us fewer patients had a detected recurrence on trial. It doesn't yet tell us whether that translates into more patients being alive, which is the actual outcome that would justify exposing already-cured patients to this risk.
I don't think either of you is wrong, and I don't think the oncology literature has settled this yet either. This is close enough, and the two sides different enough in kind — a real chance of benefit against a real chance of irreversible harm for no benefit at all — that I'd lay it out for her exactly that way and let her decide, rather than either of you steering her toward your own default.
Agreed: the tradeoff would be presented to her directly and plainly — the disease-free survival benefit and its evidentiary limits on one side, the irreversible toxicity risk and the real chance she is already cured on the other — with her own answer deciding what happens next, not either physician's recommendation.
Not agreed between the two oncologists, and stated as such rather than resolved:
Would recommend starting pembrolizumab if asked directly, given the robust disease-free survival signal.
Would recommend surveillance alone if asked directly, given the unconfirmed survival benefit and irreversible toxicity risk.
Both agreed neither leaning would be voiced to her as a recommendation — only the balanced tradeoff itself, so her decision reflects her own risk tolerance rather than whichever physician happened to speak last.