A provocative reassessment of the PROTEUS study is challenging what it means to call a cancer trial “positive” when most patients do not achieve a complete disappearance of detectable disease. In a 2026 perspective published in Nature Reviews Clinical Oncology, Daniel E. Spratt argues that the language used to describe success in oncology can obscure the experience of the majority of patients, particularly when a trial’s most celebrated endpoint is reached by only a minority.
The debate centers on a fundamental problem in modern cancer research: the difference between improving a measurable biological marker and delivering a durable clinical benefit. In diseases such as high-risk prostate cancer, treatment may combine systemic therapy with surgery or radiotherapy, aiming to eliminate both visible tumors and microscopic deposits that cannot be detected by conventional imaging. Yet even when scans appear clear and a surgical specimen contains no obvious tumor, residual malignant cells may remain elsewhere in the body.
“No evidence of disease,” often abbreviated NED, is therefore a powerful but limited concept. It generally means that available tests have failed to identify cancer at a particular point in time. It does not necessarily mean that every cancer cell has been eradicated. The distinction is especially important in cancers with a substantial risk of recurrence, where microscopic disease can remain dormant for years before becoming clinically apparent.
PROTEUS has attracted attention because it evaluates an intensified treatment strategy designed to improve outcomes before definitive local treatment. The scientific rationale behind such approaches is straightforward: exposing cancer to systemic therapy while the disease burden is still relatively limited may eliminate microscopic metastases, shrink the primary tumor and reveal how sensitive the cancer is to treatment. The approach also creates an opportunity to study tumor tissue obtained after treatment, allowing researchers to examine whether therapy has produced a pathological response.
Pathological response is often treated as an early window into a patient’s future. A complete pathological response, in which no viable cancer is found in the tissue removed during surgery, can indicate that a tumor was highly vulnerable to treatment. But it is not identical to cure. A patient may have no detectable tumor in the surgical specimen and still later experience biochemical relapse or metastatic progression. Conversely, a patient with residual tumor may remain disease-free for many years. The relationship between a short-term pathological result and long-term survival is therefore informative, but imperfect.
Spratt’s critique highlights the danger of allowing an impressive minority result to define the overall success of a trial. A study can achieve a statistically significant improvement in a prespecified endpoint while leaving most participants without the outcome that captures the public imagination. Statistical significance indicates that an observed difference is unlikely to be explained by random variation alone; it does not establish that the benefit is large, durable or meaningful for every patient.
The issue also exposes the limitations of surrogate endpoints. A surrogate endpoint is a laboratory, imaging or pathological measurement used as a substitute for outcomes that matter directly to patients, such as freedom from metastasis, avoidance of additional treatment, quality of life or overall survival. Surrogates can accelerate clinical research, but they must be validated. A treatment that improves a pathological measure without extending life or preventing complications may represent biological activity rather than a decisive therapeutic advance.
The PROTEUS discussion arrives as oncology increasingly uses combination treatments, molecular monitoring and increasingly sensitive imaging. Tests capable of detecting tiny quantities of tumor-derived DNA may identify disease that would previously have been classified as undetectable. This creates a moving target: as diagnostic tools become more sensitive, fewer patients may qualify as truly disease-free, even if their long-term outcomes improve. The definition of success can change with the technology used to measure it.
The larger lesson is not that PROTEUS or similar trials are failures. Rather, the study raises a question about how results should be communicated and judged. Researchers may need to report the proportion of patients who achieve complete response, the duration of that response and the number who remain free of metastasis over time, rather than relying on a single headline endpoint. Patients also need clear information about absolute benefit, treatment-related side effects and the likelihood that therapy will alter the course of their disease.
For cancer science, the most meaningful definition of a positive trial may be more demanding than a favorable statistical result. It should connect tumor biology with outcomes that patients can feel and understand: living longer, avoiding metastatic disease, preserving urinary and sexual function, reducing treatment burden and maintaining quality of life. PROTEUS has become a case study in why the vocabulary of success matters. When most patients never reach no evidence of disease, the central question is not simply whether a trial was positive, but how much it changed the lives of the people enrolled.
Subject of Research: Prostate cancer treatment outcomes and the interpretation of success in the PROTEUS trial
Article Title: How positive was PROTEUS? Reconsidering success when most patients never reach no evidence of disease.
Article References: Spratt, D.E. How positive was PROTEUS? Reconsidering success when most patients never reach no evidence of disease. Nat Rev Clin Oncol (2026). https://doi.org/10.1038/s41571-026-01193-4
Image Credits: AI Generated
DOI: 10.1038/s41571-026-01193-4
Keywords: PROTEUS trial, prostate cancer, no evidence of disease, pathological response, cancer clinical trials, surrogate endpoints, oncology, treatment outcomes, metastasis-free survival
