The study team reviewed US FDA oncology approvals from the launch of pembrolizumab in 2014 through early 2025 and categorized agents as immunotherapy or non-immunotherapy. They reported 143 immunotherapy approvals and 211 non-immunotherapy approvals in that period. Short sentences state the core finding plainly. The cohort makeup therefore favored non‑immunotherapy approvals numerically, while immunotherapies formed a substantial and growing subset.
Authors analyzed time-to-approval and trial scale across the decade. Early in the period, immunotherapies reached approval faster than non-immunotherapies; however, the gap narrowed over time. By 2024/2025 the reported median approval duration was 2079 days for immunotherapies and 2357 days for non-immunotherapies. Trial scale contributed to these timelines: the median trial size for immunotherapy approvals was 461 participants, and for non-immunotherapies it was 698 participants, investigators reported. These numbers suggest trial size has influenced development duration and regulatory review timing. The manuscript did not provide granular year-by-year sampling for every approval; readers should consult the underlying dataset for more detailed temporal breakdowns, the authors noted.
The authors documented a broad shift toward surrogate endpoints as primary measures supporting approval. They reported that surrogate endpoints have “overwhelmed” overall survival in frequency as the primary endpoint. Quantitative comparisons followed: for non-immunotherapies a median of 60% of the benefit predicted from progression/disease/recurrence/event-free survivals translated to overall survival. For immunotherapies the cohort showed a 99% translation of surrogate benefit to overall survival. Investigators framed this as a stark difference in surrogate validity between the two classes. The preprint did not provide additional statistical modeling details or confidence intervals in the abstract; those methodological specifics were not reported there and would require consulting the full manuscript or data files, the authors stated.
The team compared the US FDA approvals with those of the EMA and concluded the two agencies’ portfolios were very similar overall. However, the authors noted that the EMA appears generally more conservative or skeptical in granting approvals. The preprint provided this comparative statement without extensive tabulation in the abstract; investigators signaled that some therapies approved by the FDA were either rejected by or pending at the EMA, which informed their decision to present the primary analysis using FDA approvals.
Investigators cited the FDA Oncology/Hematologic Malignancies Approval Notifications and ClinicalTrials.gov as the primary information sources. Data used in the analyses are reported to be available upon reasonable request to the authors. The preprint carries standard author declarations: no competing interests were declared, appropriate ethical guidelines and trial registrations were confirmed, and the authors provided contact information for correspondence. The paper remains a preprint and has not undergone peer review; authors and the posting platform reminded readers that the findings should not be used to guide clinical practice without further validation.
Authors identified several open issues. Trial size and trial duration were singled out as problems needing attention—specifically, investigators argued these parameters should be meaningfully reduced to accelerate development. They also questioned the validity and defense of surrogate endpoints given the discordant translation rates to overall survival between immunotherapies and non-immunotherapies. The study team noted that regulatory differences between the FDA and EMA warrant further scrutiny but did not outline prescriptive policy steps in the abstract. The manuscript ends on these concrete limitations and unanswered questions rather than a generic wrap-up. Readers seeking methodological details, per‑approval breakdowns, or statistical testing should consult the full text and available data files; those specifics were not provided in the abstract of the preprint.