Belzutifan’s early disease-free survival gain weakens after treatment stops in kidney cancer trials
Adding belzutifan to adjuvant pembrolizumab improved disease-free survival in high-risk renal cell carcinoma, but the benefit did not behave like a straight line once the planned treatment period ended. A reconstructed individual patient data analysis of phase 3 adjuvant kidney cancer trials found that the hazard advantage was strongest before month 12, then attenuated afterward, raising a practical question for clinicians: how much of the observed win persists once therapy is withdrawn?
The study centered on LITESPARK-022, where central review showed a 24-month restricted mean survival time difference of 1.34 months, with the belzutifan arm still ahead after month 12. By the investigators’ reconstruction, the incremental gain after month 12 amounted to 0.74 months per year. In plain clinical terms, the curve kept separating, but far less forcefully than it had during active treatment.
A sharper early hazard reduction, then a flatter post-treatment curve
Before month 12 in LITESPARK-022, the hazard ratio for disease-free survival was 0.55 with central review, indicating a substantial reduction in recurrence or death during the treatment window. After month 12, the hazard ratio shifted to 1.16, a reversal that suggests the relative protection was no longer clearly sustained once the protocol-defined treatment period had ended. The post-to-pre hazard ratio ratio reached 2.10, an unusually stark sign of temporal decay in treatment effect.
A single summary hazard ratio would have blurred this pattern. Over the 12- to 30-month interval, the absolute disease-free survival difference changed by −3.7 percentage points with central review. With investigator assessment, the same interval showed a 0.6-point change, which points to another problem in survivorship analysis: the estimate can move depending on who reads the scan and when the event is called.
The biological logic is not mysterious. Belzutifan inhibits HIF-2α, a pathway tied to the hypoxia signaling machinery that drives clear-cell renal cell carcinoma. During active suppression, the drug may delay microscopic residual disease from re-expanding. Once exposure ends, any residual clones that escaped treatment pressure can resume growth, and the curve can flatten or drift back toward the control arm. That pattern fits a therapy that suppresses disease rather than eradicating every surviving tumor cell.
Similar time-dependent patterns appeared across other adjuvant renal cancer trials
The analysis did not examine LITESPARK-022 in isolation. Investigators reconstructed survival data from ten phase 3 adjuvant renal cell carcinoma trials and calculated restricted mean survival time differences, hazard ratios before and after planned treatment ended, and disease-free survival differences within each trial.
Comparable post-to-pre hazard ratio ratios appeared in other studies. In KEYNOTE-564 at its first interim analysis, the ratio was 2.02, yet the benefit persisted. In SORCE and PROTECT, the ratios were 2.09 and 2.59, respectively, and the benefit faded. Those parallel trajectories matter because they show how a treatment can produce a robust early separation while still leaving open the question of durability once the intervention is over.
Here, the difference between a sustained effect and a fading one appears to depend on the trial’s internal timing as much as on the drug itself. Post-treatment follow-up in adjuvant oncology is often where the strongest assumptions get tested, because the patient is no longer receiving the drug, yet the endpoint continues to accumulate. When recurrence risk is front-loaded during the first year, a regimen may look especially effective early and less convincing later, even if its initial impact was clinically meaningful.
Why restricted mean survival time tells a different story than a single hazard ratio
Restricted mean survival time offers a more concrete way to measure benefit over a fixed horizon. Instead of compressing the entire curve into one proportional hazard estimate, it asks how much additional disease-free time the treatment arm actually buys over a prespecified period. In LITESPARK-022, that 24-month gain was 1.34 months by central review, which is modest but tangible.
The central message is not that belzutifan failed. It is that the magnitude and persistence of benefit are not the same thing. A therapy can reduce early events, lengthen mean time without recurrence, and still lose relative momentum after treatment ends. For oncologists counseling patients after nephrectomy, that distinction affects how one frames expected durability, especially when balancing benefit against the toxicity profile of combination adjuvant therapy.
The patient summary in the preprint is blunt about this tradeoff: the addition of belzutifan increased the average time without cancer returning by about 6 weeks over 2 years, but caused more serious side effects. That kind of signal forces a bedside calculation, because the value of adjuvant treatment in renal cell carcinoma hinges not just on whether the curve bends, but on how long it stays bent.
Assessment method and follow-up length shape the clinical interpretation
The analysis depends on reconstructed individual patient data derived from published Kaplan-Meier curves and reported summary statistics. That approach is increasingly useful when trial-level publications do not yet provide mature patient-level data, but it still carries a structural limitation: reconstructed curves cannot fully replace original time-to-event records, and they are only as faithful as the published graphs allow.
Follow-up was also short. In adjuvant renal cell carcinoma, where late recurrences can still emerge, brief observation can exaggerate confidence in either direction. An early advantage may shrink with more time, or a modest early gap may widen if late protection persists. This preprint cannot answer that question definitively, because the available follow-up does not extend far enough to settle long-term durability.
The practical takeaway is narrow and actionable. Belzutifan appears to contribute an early disease-free survival advantage when layered onto adjuvant pembrolizumab, but the post-treatment trajectory is less secure than a single hazard ratio would suggest. For now, single-HR projections should be avoided, because they flatten the very feature that matters most here: whether the benefit survives beyond the planned treatment period.
What clinicians should watch as more data mature
If longer follow-up confirms the post-month-12 attenuation seen here, adjuvant strategy in renal cell carcinoma may need to be judged less by initial curve separation and more by durability after therapy stops. If the benefit proves sustained in later readouts, the early hazard reduction will look more reassuring than it does now. Either way, the current analysis argues against overinterpreting a short-term proportional hazard estimate as proof of lasting disease control.
For patients, the question is simple and immediate: does the extra time without recurrence hold after treatment ends, or does it mostly buy a temporary delay? This preprint cannot close that gap, and the limited follow-up leaves the answer unresolved.