Clinical endpoints, specifically overall survival (OS), are crucial to assess cancer therapies. They are costly and time-consuming. Surrogate endpoints (SEPs), such as pathologic complete response (pCR), defined as the absence of viable tumor after treatment, have gained increased prominence, especially following the Food and Drug Administration (FDA) approval of pCR in 2012. While pCR patients showed survival links in certain cancers like rectal and breast cancer, trial-level associations remain weak and raise concerns about the validity of pCR across various cancer types. This has resulted in the regulatory authorization of expensive medicines without a survival benefit.
A study published in JAMA Network Open conducted a systematic review and meta-analysis to evaluate the trial-level association of pCR with survival outcomes in rectal cancer. This study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. They have used Cochrane, EMBASE, and PubMed databases till January 3, 2024, to identify randomized controlled trials (RCTs) on rectal cancer that reported on neoadjuvant therapies, disease-free survival (DFS), OS, and pCR. Two reviewers have selected studies and extracted data. Weighted linear regression was used to analyse the association between survival outcome and pCR. Statistical heterogeneity was evaluated using random-effects meta-analysis, I² statistics, and χ² tests. The risk of bias was assessed using the Cochrane tool, along with the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach. Sensitivity analyses were also performed to ensure the certainty of the evidence.
The meta-analysis included 25 RCTs with a total of 11,882 rectal cancer patients, selected from an initial pool of 3363 studies. All included studies defined pCR as the absence of lymph node metastasis or residual tumor. This study showed no significant relationship between OS and pCR, pr, and DFS. Weighted regression analysis revealed that pCR was not linked with OS (95% confidence interval [CI] −0.98 – 1.71, β = 0.37, p = 0.57) or DFS (95% CI, −2.55 – 0.87, β = −0.84, p = 0.32). Sensitivity analyses excluded studies with high risk of bias, non-curative resections, nonproportional hazard ratios, and significant heterogeneity. Additionally, eliminating historical studies by using radiation therapy alone did not modify the outcomes. Subgroup analyses were done to reduce the biases, but no link between OS and pCR. The risk of bias was low to mild in the studies, and publication bias was mitigated by using the Egger test and funnel plots. The GRADE assessment evaluated the confidence of evidence as high. These results indicate that pCR must not be used as a valid surrogate endpoint for the long-term survival in rectal cancer treated with neoadjuvant therapies.
This discovery is consistent with problems raised in breast cancer research, where pCR is primarily used for expedited approvals due to insufficient trial-level validation. The absence of association may be due to pCR corresponding only to the primary tumor response, rather than micrometastatic disease, and could be confounded by subsequent adjuvant therapy. Although complete neoadjuvant treatment is now widely used in rectal cancer, few complete neoadjuvant treatment studies have permitted comprehensive subgroup analysis.
This study’s limitations include variability in treatment methods, the inclusion of some patients who did not undergo surgery, inconsistent follow-up periods, and difficulty assessing hazard ratio proportionality across all trials. Despite tripling the number of trials compared to earlier studies, the results remained unchanged, indicating strong evidence against using pCR as a proxy for survival in rectal cancer.
This comprehensive review and meta-analysis of neoadjuvant therapy in rectal cancer revealed no trial-level correlation between pCR and survival, which indicates that the use of pCR as the surrogate endpoint for survival must be reconsidered and implemented with caution. Further studies should explore patient-level data and identify additional biomarkers to improve the accuracy of survival prediction.
Reference: Sugumar K, Lie JJ, Stucky C, et al. Pathologic Complete Response and Survival in Rectal Cancer: A Systematic Review and Meta-Analysis. JAMA Netw Open. 2025;8(7):e2521197. doi:10.1001/jamanetworkopen.2025.21197
