JCO Erratum Corrects Survival Figures in Ewing Sarcoma ctDNA Study
The Journal of Clinical Oncology issued an erratum on October 14, 2026, to correct numeric and labeling errors in a study that used circulating tumor DNA to define risk in patients with Ewing sarcoma. The corrected paper is the report from the Children’s Oncology Group and the LEOPARD Study, originally published online September 15, 2026.
Two of the changes alter survival data in clinically meaningful ways. One revision corrects the overall survival figure for patients with localized disease and pretreatment ctDNA burden ≥0.5%, changing the 2-year OS from 38% to 93% while leaving the comparator group at 100%. Another correction changes a mislabeled survival endpoint in the metastatic cohort, replacing an incorrect reference to EFS with OS in the sentence describing inferior 2-year outcomes among patients with residual on-treatment ctDNA.
A third correction adjusts the baseline mutation analysis, where the tissue-and-plasma subgroup size was tied to assessment of STAG2 and TP53 alterations in patients with localized disease and ctDNA burden above 0.5%.
The corrections are not cosmetic. In a biomarker-driven sarcoma study, survival percentages guide how investigators interpret risk, biological aggressiveness, and the potential value of intensifying therapy or refining trial stratification. A mistaken overall survival estimate can distort the apparent separation between risk groups and blur the clinical meaning of a ctDNA threshold that is intended to sort patients by prognosis.
Why a ctDNA Threshold Carries Clinical Weight
The corrected analysis centers on a pretreatment ctDNA burden cutoff of 0.5%. In localized Ewing sarcoma, the original article linked ctDNA burden ≥0.5% with inferior event-free survival and overall survival compared with a lower-burden group. Once the OS figure is corrected to 93% in the higher-burden group, the direction of risk remains unfavorable for the ctDNA-positive cohort, but the magnitude of the difference becomes far less extreme than the erroneous 38% figure suggested.
That matters because ctDNA is being developed as a real-time molecular readout of tumor burden and biologic behavior. In a disease such as Ewing sarcoma, where microscopic dissemination may exist before imaging declares progression, ctDNA can function as a proxy for disease activity that is not yet clinically visible. A biomarker that divides patients into risk strata must be numerically exact; otherwise, the threshold itself may appear more or less definitive than the underlying data justify.
One sentence in the erratum changes the endpoint language in the metastatic cohort from EFS to OS. In oncology reporting, that distinction is fundamental. Event-free survival captures relapse, progression, second malignancy, or death, while overall survival reflects death alone. A mislabeled endpoint can lead clinicians to assume a different failure pattern than the one actually measured, especially in a setting where residual on-treatment ctDNA is being used to identify patients at high risk of relapse.
Residual ctDNA and the Biology of Persistent Risk
The corrected metastatic-disease passage still conveys the same clinical message: patients with residual on-treatment ctDNA had inferior outcomes, including worse 2-year overall survival. Persistently detectable tumor DNA during treatment implies that therapy has not fully suppressed the malignant clone. Biologically, that persistent signal can reflect resistant subclones, incomplete cytoreduction, or ongoing shedding from active disease sites that remain outside the reach of conventional response assessment.
In Ewing sarcoma, where treatment response often depends on early control of systemic disease, residual ctDNA offers a molecular snapshot of incomplete eradication. A rising or persistent signal during therapy can be more informative than static anatomic imaging, because DNA released into plasma tracks with viable tumor burden. The corrected survival endpoint strengthens the interpretive framework for that idea: the concern is not merely that disease has not disappeared radiographically, but that the molecular residue of the tumor remains detectable while treatment is underway.
One endpoint label can change how an entire dataset is read.
Baseline Mutation Analysis and Risk Stratification
The erratum also corrects a sentence in the analysis of baseline STAG2 and TP53 alterations. The subgroup included patients with localized disease and available tissue or a baseline plasma sample with ctDNA burden >0.5%, totaling 119 patients. Even without adding any new results beyond the correction itself, the revised wording matters because mutation analyses in Ewing sarcoma sit at the intersection of tumor biology and risk assignment.
STAG2 and TP53 are among the alterations often discussed in aggressive cancer biology because they may reflect genomic instability and impaired tumor control. When paired with ctDNA burden, such mutations can help define a disease state that is not only anatomically localized but molecularly high-risk. Precision in describing the cohort is essential, because a mislabeled sample size can obscure how much confidence to place in any association between mutation status and survival.
What the Correction Means for Readers and Researchers
For practicing oncologists, the erratum does not change the core premise that ctDNA profiling can help classify risk in Ewing sarcoma. It does, however, change how sharply the survival curves should be interpreted. The corrected 2-year overall survival of 93% in the localized, higher-ctDNA group preserves the prognostic signal while avoiding a dramatic underestimation of survival that would have overstated the separation between groups.
For investigators, the corrections are a reminder that biomarker studies depend on exactness at every level: assay threshold, endpoint definition, subgroup designation, and numerical reporting. ctDNA has clinical promise precisely because it can convert circulating tumor fragments into a measurable marker of residual disease and relapse risk. If the numbers are wrong, the biology is harder to trust, and the trial design built on that biology becomes less reliable.
This erratum leaves the broader scientific question intact. ctDNA appears to stratify risk in Ewing sarcoma, and residual molecular disease during treatment remains a worrisome sign. The corrected article now presents those signals with the proper survival values and endpoint language, which is the minimum requirement for any biomarker meant to influence prognostic assessment or future trial design.