Rhabdomyosarcoma is one of the most aggressive forms of cancer in children and adolescents. Even with intensive treatment, the chances of a cure are low in metastatic cases. The hope that an allogeneic stem cell transplant might offer these children a survival advantage was scientifically plausible – yet proved to be unsustainable. In this interview, Uwe Thiel and Sebastian J. Schober explain what this means, how the researchers reacted to this finding, and why they nevertheless regard it as a success.

“A 20 per cent transplant-associated mortality rate – that is clear-cut data.”

When new approaches are subject to critical discussion: how do you tell whether criticism helps you move forward or tends to hold you back?

This trend was already becoming apparent in the first systematic analysis of survival data. We viewed this as a positive development, as it allowed us to derive a clear clinical recommendation: no more toxic treatment without measurable benefit. A 20 per cent transplant-associated mortality rate with no survival benefit – that is unequivocal evidence.

Research without risk – that sounds paradoxical. Can you explain why research that cannot also fail fails to fulfil its actual purpose?

If we only test hypotheses where we can predict the outcome, we won’t make any progress. Particularly in paediatric oncology, where patient numbers are low and prognoses are often poor, we must have the courage to ask uncomfortable questions too. The hypothesis of a graft-versus-tumour effect was biologically plausible. The fact that it was not confirmed in our cohort now protects future patients from a gruelling course of treatment. This is not a failure, but precisely the purpose of research.

What, for you personally, was the greatest risk of this project – from a scientific, clinical or professional perspective?

The scientific risk was calculable. The practical risk lay in the multi-centre collaboration across several European countries, which requires trust and tact. As far as one’s career is concerned, the pressure to publish positive results is very real. You have to separate that from your ego and remember the original remit. If you don’t publish negative results, you’re wasting other research groups’ time.

“Around half of all published biomedical results cannot be reproduced.”

Were you concerned that a negative result would be more difficult to publish?

No, reputable journals also publish negative results. Three international reviewers, including one from the Memorial Sloan Kettering Cancer Centre, immediately recognised the relevance of our work. However, this is not a given, as systematic publication bias is well documented. Around half of all published biomedical results cannot be replicated. At Amgen, for example, only six out of 53 preclinical cancer studies could be confirmed. As long as the incentive system rewards quantity rather than reproducibility, this will remain a structural problem.

In your opinion, when is a research project actually considered a failure – and when is it not?

A project has failed if the pressure to publish leads to a situation where one’s own results can no longer be viewed objectively. Or if, as a community, we accept that a significant proportion of our results are not reproducible. The estimated annual cost of this in the US alone runs into the tens of billions. Ioannidis described this as early as 2005. Twenty years on, little has changed structurally. By contrast, a project that provides a clear answer to a question is not a failure, regardless of the result.

“The door remains open for combined approaches.”

Your findings suggest a clear recommendation against stem cell transplants for certain patient groups. What responsibility do you see in drawing such a conclusion?

We see this responsibility on three levels. Firstly: transparency. We have published all the data in full, including the high transplant-associated mortality rate. Secondly: methodological rigour. The matched-pair analysis with non-transplanted control patients was crucial in demonstrating the lack of added benefit. And thirdly: consistency. Our conclusion is clear. At the same time, a distinction must be made. A prospective study involving additional donor lymphocyte infusions could reassess the issue. We have described a patient with an Oberlin score of 4, who survived at least 5 years without recurrence following transplantation and DLI. This individual case demonstrates that alloreactive immune mechanisms could be effective under certain conditions. The door remains open for combination approaches.

How have you translated the findings from this project into clinical practice? What specific changes have there been for patients?

We no longer perform transplants on these patients. There are fewer hospital stays involving, in some cases, severe toxicity without clinical benefit – that is a concrete, positive outcome. At the same time, the realisation that alloreactive donor cells cannot control solid tumours on their own has changed the direction of our research. We are now conducting research into a combination of oncolytic viruses, CDK4/6 inhibitors and TCR/CAR-T cells, which specifically alter the tumour microenvironment. The negative result has therefore opened up a new avenue of research.

“Funding bodies that recognise the value of a properly conducted negative study act as a crucial corrective.”

How important is it to communicate negative results openly and transparently, even if they temper hopes or disappoint expectations?

It is essential. Failure to publish negative results allows other research groups to repeat the same mistakes. This costs time and resources that could be spent on other projects. Furthermore, it contributes to the reproducibility crisis, which undermines trust in biomedical research. Funding bodies such as the Wilhelm Sander Foundation, which recognise the value of a properly conducted negative study, therefore act as a crucial corrective. After all, it is ultimately the funding conditions that determine whether researchers have the courage to ask uncomfortable questions and publish uncomfortable answers.