A tumor response, time without progression and survival are different outcomes. A favorable result on one does not automatically establish a favorable result on the others. To understand a cancer-trial headline, identify the endpoint, patient population, comparison and follow-up before interpreting the percentage. 123
This page supplies a reading method, not a ranking of cancer treatments. Evidence from a different subtype or treatment setting should not be transferred to a person's care because a headline uses the same broad cancer name.
Four endpoints that should stay separate
| Term | What it describes | What it does not automatically prove |
|---|---|---|
| Objective response rate, or ORR | The proportion whose cancer shrinks or disappears according to the stated response assessment | That the same proportion is cured or lives longer |
| Progression-free survival, or PFS | Time alive without the disease worsening | That all tumors disappeared or overall survival improved |
| Disease-free survival, or DFS | Time after treatment without signs of the cancer returning, in the relevant study setting | A lifetime guarantee that recurrence will not occur |
| Overall survival, or OS | Time people remain alive from the study's specified starting point | That an observed difference was caused by treatment without an appropriate study comparison |
These definitions follow NCI terminology. Individual trial protocols define how events, assessment and analysis are handled, so the endpoint name is the beginning of interpretation rather than the complete methods section. 1243
Keep response and durability apart
A response-rate statement answers how many people met a response definition. It does not, without additional results, describe how long those responses lasted or what happened to people who did not respond.
For example, in a fictional trial of 100 assessed participants, 30 meet the specified response criterion. The response rate is 30%. That does not mean every participant's tumor shrank by 30%, or that 30 people were permanently cured.
Write down both the numerator and denominator: “30 of 100 assessed participants met criterion X.” Then ask whether the source reports duration, later events, missing assessments and the trial's comparison group. These are reading questions, not a claim that a particular real trial omitted them.
Compare at the same time and in the same setting
Imagine a second fictional statement: 70 of 100 participants are alive at two years, with complete follow-up for this simplified example. That is a different observation from a response rate. It cannot be compared directly with another study's 70% response figure as though the values measure the same benefit.
Likewise, a result after surgery intended to prevent recurrence answers a different question from a trial treating measurable advanced disease. Before placing two percentages side by side, record whether the studies address the same treatment stage, population and outcome.
A clear comparison may conclude, “Not directly comparable from the information supplied.” That is a more useful result than a confident winner based on mismatched percentages.
Do not translate a relative number into an absolute promise
Suppose an original arithmetic example has an event risk of 20% in one group and 15% in another at a stated time with complete follow-up. The absolute difference is five percentage points; the relative reduction in this simplified risk comparison is 25%.
That is not permission to interpret every reported hazard ratio as the same calculation. Use the study's actual measure and methods. The absolute-versus-relative benefit guide explains why the starting risk and follow-up belong beside a relative figure.
When the source supplies only a headline relative result, leave the absolute benefit as unknown until the relevant data are available. Do not manufacture a number from an unrelated earlier trial.
Record the study's planned question and its reported result
Ask whether the endpoint was identified as primary or secondary, whether the source says the planned analysis was met, and whether you have the full report or only an announcement. Preserve the source type in the record.
A press release saying a primary endpoint was met is not the same document as a peer-reviewed paper with detailed outcomes. Nor is a conference report automatically an independent replication of earlier findings in the same participants. The source-type guide explains that distinction.
Use the endpoint-comparison sheet to keep population, endpoint, comparison, time and source together. No clinical ranking is calculated.
Add harms and burden before calling a result useful
Even a clearly reported benefit does not answer every treatment question. Locate the source's safety outcomes, follow-up, discontinuations and treatment burden, and identify which information remains unavailable. The purpose is to understand the complete reported evidence, not to replace the treating team's judgment.
Apply this method to the site's cancer-vaccine evidence collection. An immune response, a tumor response, less recurrence and longer survival are all meaningful research questions—but they are not interchangeable announcements of a cure.