We will not have to wait—building regulatory confidence in early endpoints for high-risk breast cancer

CamRelief is a well-conducted phase III neoadjuvant trial in stage II–III triple-negative breast cancer (TNBC) that shows a class-consistent benefit: adding the programmed cell death protein 1 (PD-1) inhibitor camrelizumab to a standard dose dense taxane-platinum-AC chemotherapy backbone increased pathological complete response (pCR) by ~12 points (56.8% vs. 44.7%) (1). This is the fundamental paradigm of neoadjuvant trials in breast cancer, to provide near-term, easily measurable endpoints of treatment response, with later verification of survival outcomes as they mature. Two JAMA pieces accompany the report, stating “we will have to wait” for survival results to mature to determine if camrelizumab provides benefit, as well as calling for the more granular residual cancer burden (RCB) results with surgical management detail (2,3). We agree with the careful discussion of the central facts. However, in the context of evidence for anthracycline-based chemoimmunotherapy provided by Keynote-522, IMpassion031, and GeparNuevo, the more salient issue is whether every new molecule in the same class requires a full superiority trial powered for 3–5-year survival outcomes before patients can access treatment. This agent, developed and trialed in China, shows extensibility of PD-1 inhibition to a different population and supports broad applicability of the class. Agents in the same class have already shown that pCR is a reliable predictor of survival benefit in this setting. In general, we believe patients should not have to wait for access to promising agents when a benefit is reasonably likely to exist. We will focus the rest of the commentary on the concept of an accelerated pathway to approval based on high quality early-endpoints, as neoadjuvant design success in accelerating access and knowledge turns hinges on both regulatory and clinician acceptance of early endpoints. We highlight three practical points.

What is an optimal early endpoint?

Early endpoints that reliably identify patients achieving durable remission shorten the time between treatment discovery and access, shift the industry incentive from long term market share to continual improvement, and while confirmatory survival follow up is critical, regulators do not need to wait to grant conditional approval through an accelerated pathway when such reliability has been established. In high-risk breast cancer, cure is achieved by elimination of micro-metastatic disease with systemic therapies, while surgery provides a pathological assessment of in-vivo treatment response. This neoadjuvant paradigm targeting a clinical cure is here to stay and likely to expand across solid tumor management standard of care.

It is in this context that pCR is a natural early endpoint for neoadjuvant therapy in TNBC, where pathological elimination of disease in an aggressive disease population is a conservative clinical target that represents a likely cure at the patient level. Many studies have demonstrated that pCR is an excellent prognostic marker for individual patients, yet as a regulatory endpoint, pCR is also being tasked with determining whether a treatment is favorable at the population level based on differences in pCR rates vs. control. While recurrence rates in pCR are consistently low, non-pCR recurrence risk is inherently heterogeneous, and population survival differences will depend on outcomes of the subset of patients with residual disease on surgical pathology. Patients with low levels of residual disease may be cured by surgery or local therapies, while recurrence rates are much higher for higher values of RCB, so while there is heterogeneity of risk, it is not unpredictable. Pooled analyses have found a robust log-linear increase in hazard of recurrence as RCB increases (4). The implication is that a treatment that only increases pCR in low risk patients may not show a survival benefit, while an effective treatment can shift the overall risk of recurrence a great deal by reducing the overall disease burden compared to controls, even if the change in rate of pCR is relatively modest (Figure 1). Reports of the full RCB score here would provide response and prognostic granularity that can be crucial for downstream interpretation, but as is, CamRelief appears to be powered and designed to provide increased certainty of both the benefit of PD-1 inhibitor immunotherapy, and another example of pCR predicting a recurrence free benefit. As with any early endpoint, initial results should be followed with confirmatory survival investigations to verify this benefit and prevent any unexposed safety concerns, without denying patients access to promising agents. Drugs in the same class, that show the same improvement in early endpoint should not need the same survival delay as a first in class agent.

Figure 1 Density plots of residual cancer burden in an experimental vs. control arm of a TNBC cohort. pCR rate depicted by bar plot. Log-hazard of distant recurrence free survival relative to pCR from a Cox regression is plotted indicating a log-linear increase with increasing residual cancer burden. Mean RCB per treatment group is indicated by vertical dashed lines. pCR, pathological complete response; RCB, residual cancer burden; TNBC, triple negative breast cancer.

Confidence also hinges on rigor: unlike primary GeparNuevo analysis, pCR in CamRelief follows the CTNeoBC standard definition for pCR as ypT0/Tis, ypN0, and like Keynote-522, patients have stage II/III disease with strong representation of nodal burden (over 70% clinical node positivity). As patients, clinicians, researchers, and regulators strive for improved early endpoints, establishing consensus around the endpoint definition and its collection are critical for validation and evaluation of treatment benefit across studies. Both pCR and RCB are examples of established pathology measures that should be recorded in all neoadjuvant trials as standard of care (5). We believe that improved early endpoints are well within reach, with RCB providing an initial candidate over pCR. RCB is prognostic at the patient level and will likely discriminate treatment effectiveness once adequate meta-analytic evidence is accrued, an effort that is helped by a rigorous definition and precise, reliable collection across trials. Meanwhile FDA guidance indicates that regulators are open to supporting more sophisticated approaches of hierarchical testing and multicomponent measures which could include imaging or liquid biopsy measures under development and lead to rapid improvements in early treatment effect estimation (6). Advances in regulatory statistical methods including modern Bayesian techniques and causal inference leveraging modern computing power are also possible directions for development. For the time being, patients hoping to try camrelizumab will have to wait for 3-year event-free survival (EFS) or distant recurrence free survival to convince regulators, highlighting the importance of an early endpoint that produces the confidence and buy-in to drive progress in the field with the urgency required.

Is the appropriate population selected for the trial?

Early endpoints work best and are most appropriate in the populations where accelerated approval pathways are meant to help. Disease-risk enrichment (TNBC, stage II–III) both aligns with unmet-need intent where a clinical trial has a favorable risk-benefit tradeoff and ensures sufficient events for confirmatory recurrence free survival. Responder enrichment should also be prespecified and powered [PD-L1, tumor-infiltrating lymphocytes (TILs), and immune gene signatures], so we don’t average away heterogeneity; the CamRelief population largely gets the risk piece right, although future trials and more granular data collection can deepen the biomarker layer prospectively.

Have we tested a strategy for therapy de-escalation based on response?

One-year adjuvant PD-1 and highly intensive chemo should not be permanent defaults. Response-aware trial design is crucial to accelerate knowledge turns and generate patient benefit both on trial and for future trial designs targeting remaining uncertainty. If we believe that pCR is associated with an excellent outcome, then our trial designs should reflect that. When a clinical cure is achieved, patients should be spared the toxicity of extended therapy, and at the very least that option must be studied. Importantly, the I-SPY 2 TRIAL showed a significant increase in pCR with only 4 cycles of a PD-1 inhibitor and the EFS benefit was maintained without adjuvant PD-1 (7). In Keynote-522, those patients on the control arm who achieved a pCR had excellent outcomes, identical to those in the experimental arm who achieved a pCR but also received a year of PD-1 therapy (8,9). For patients achieving pCR or minimal residual disease, continuing a year of adjuvant immunotherapy may add toxicity without measurable benefit. Designing trials that formally test de-escalation in these responders honors the principle of “first, do no harm” while embedding the goal of clinical cure. An early endpoint that reliably identifies patients achieving durable remission should allow for both accelerated approval and de-escalation of therapy, but our trial designs must ask the question: Is the dose and duration of therapy necessary to achieve the desired outcome? These questions remain unanswered awaiting trials of early responders who should be allowed to avoid the toxicity of additional chemotherapy, non-responders who could be re-randomized given the diminishing returns of additional chemotherapy cycles, and the assessment of non-inferiority of discontinuation of adjuvant immunotherapy for patients achieving pCR or RCB1.

Overall, CamRelief adds another consistent pCR signal for neoadjuvant PD-1 in high-risk TNBC; and follow-up for survival will confirm a benefit. Accelerated solutions for our highest risk patients require harmonized efforts to define and establish early endpoints, enrich for high-risk disease and predicted responsiveness, and embed adaptation. If we demand mature follow-up before approving access, we are abandoning a neoadjuvant paradigm for trial designs that cures patients, speeds up knowledge acquisition, and saves lives through early access. Assessments of long-term recurrence and survival outcomes can and will occasionally find unexpected results, and our regulatory agencies must be equipped to handle these rare situations, rather than block patients from accessing promising treatments that have not yet withstood the test of time. What is necessary is early endpoints and trial designs that account for complexity of disease heterogeneity while providing both patient level cures and regulatory confidence.

Acknowledgments

None.

Footnote

Provenance and Peer Review: This article was commissioned by the editorial office, AME Clinical Trials Review. The article has undergone external peer review.

Peer Review File: Available at https://actr.amegroups.com/article/view/10.21037/actr-25-123/prf

Funding: None.

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://actr.amegroups.com/article/view/10.21037/actr-25-123/coif). L.E. reports grants or contracts from Moderna; support for attending meetings and/or travel from Blue Cross Medical Advisory Panel; and serves as an uncompensated board member of Quantum Leap Healthcare Collaborative. The other author has no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Chen L, Li H, Zhang H, et al. Camrelizumab vs Placebo in Combination With Chemotherapy as Neoadjuvant Treatment in Patients With Early or Locally Advanced Triple-Negative Breast Cancer: The CamRelief Randomized Clinical Trial. JAMA 2025;333:673-81. [Crossref] [PubMed]
2. Joensuu H. Neoadjuvant Camrelizumab for Triple-Negative Breast Cancer. JAMA 2025;333:670-2. [Crossref] [PubMed]
3. Pei YY. Camrelizumab vs Placebo in Patients With Triple-Negative Breast Cancer. JAMA 2025;333:1835. [Crossref] [PubMed]
4. Yau C, Osdoit M, van der Noordaa M, et al. Residual cancer burden after neoadjuvant chemotherapy and long-term survival outcomes in breast cancer: a multicentre pooled analysis of 5161 patients. Lancet Oncol 2022;23:149-60. [Crossref] [PubMed]
5. Litton JK, Regan MM, Pusztai L, et al. Standardized Definitions for Efficacy End Points in Neoadjuvant Breast Cancer Clinical Trials: NeoSTEEP. J Clin Oncol 2023;41:4433-42. [Crossref] [PubMed]
6. Multiple Endpoints in Clinical Trials - Guidance for Industry. Clin Trials.
7. Nanda R, Liu MC, Yau C, et al. Effect of Pembrolizumab Plus Neoadjuvant Chemotherapy on Pathologic Complete Response in Women With Early-Stage Breast Cancer: An Analysis of the Ongoing Phase 2 Adaptively Randomized I-SPY2 Trial. JAMA Oncol 2020;6:676-84. [Crossref] [PubMed]
8. Schmid P, Cortes J, Pusztai L, et al. Pembrolizumab for Early Triple-Negative Breast Cancer. N Engl J Med 2020;382:810-21. [Crossref] [PubMed]
9. Schmid P, Cortes J, Dent R, et al. Overall Survival with Pembrolizumab in Early-Stage Triple-Negative Breast Cancer. N Engl J Med 2024;391:1981-91. [Crossref] [PubMed]