The statistical question behind a positive result matters as much as the word “positive.” This melanoma trial illustrates why response activity and survival superiority should not be merged. Primary study
Trial structure and findings
The 2010 report randomized 488 patients among five dacarbazine-based regimens involving thymosin alpha-1, interferon alfa, or both. Best overall response at 12 months was the primary endpoint. Two thymosin-containing arms exceeded a prespecified response-rate benchmark. Reported overall-survival and progression-free-survival comparisons did not reach conventional statistical significance. Primary report
Our appraisal: identify the null hypothesis
Rejecting a fixed historical response threshold is not identical to showing superiority over a concurrently randomized control arm. Those analyses have different reference points. An accurate summary should specify which test supports each conclusion instead of treating every favorable comparison as the primary randomized result.
Five treatment arms also create several potential contrasts. Interpretation depends on the prespecified analysis plan: whether arms were pooled, which comparisons were primary, and how multiplicity was handled. A pooled estimate should retain the fact that its component regimens were not identical.
Tumor response and survival are different outcomes. A regimen may produce an activity signal without a sufficiently precise survival estimate. Conversely, a nonsignificant survival test does not prove identical survival. Confidence intervals, follow-up, and event counts are needed to describe the remaining uncertainty.
This is evidence about combinations studied in a particular historical treatment setting, not a stand-alone thymosin cancer-treatment claim. The accessible abstract contains an apparent confidence-level labeling irregularity; numerical survival intervals should be checked against the publisher's full report before reuse. A useful reading retains the exploratory activity signal while withholding a confirmed survival-benefit conclusion.
Frequently asked questions
How many regimens were compared?
Five.
Was the primary endpoint overall survival?
No; it was best overall response.
Is a fixed benchmark the same as a randomized control?
No.
Was survival superiority conclusively established?
No.
Does nonsignificance prove equal survival?
No.
Was this thymosin monotherapy?
No.
What needs checking before quoting survival intervals?
The full report's confidence-level labeling and analysis definitions.