Ewing Sarcoma’s EWS::FLI1 Trial Just Showed What Precision Oncology Has Been Promising for a Decade

Picture a trial team staring at a Ewing sarcoma patient cohort — a disease so rare that enrolling 30 patients qualifies as a large study, where the oncogenic driver is a chromosomal translocation so consistent it defines the tumor itself. That driver is EWS::FLI1, a fusion protein present in roughly 85% of Ewing sarcoma cases, and for years it was considered largely undruggable. The phase 1/2 trial of trabectedin combined with low-dose irinotecan, published in Nature Medicine in 2026, just changed that assumption in ways that extend well beyond sarcoma.

The study’s design is the signal worth decoding. This was not a histology-agnostic basket trial casting a wide biomarker net. It was built around a single, molecularly defined target — a transcription factor fusion — in a disease defined by that fusion. That specificity is both the scientific thesis and the operational template.

The result is a trial architecture that the rare cancer and pediatric oncology communities should study as carefully as the efficacy data itself.

The Fusion Protein as the Protocol

EWS::FLI1 arises from a translocation between chromosomes 22 and 11, fusing the EWSR1 gene to FLI1, a transcription factor in the ETS family. Because transcription factors lack the enzymatic binding pockets that small molecules typically exploit, targeting EWS::FLI1 directly has historically been treated as a dead end. The trabectedin-irinotecan combination takes an indirect route: trabectedin, a DNA minor groove binder originally derived from the sea squirt Ecteinascidia turbinata and approved in the United States and Europe for soft tissue sarcoma and ovarian cancer, disrupts EWS::FLI1 transcriptional activity by interfering with its DNA-binding program. Low-dose irinotecan adds topoisomerase I inhibition as a complementary mechanism, targeting the transcriptional dependencies that EWS::FLI1 creates.

This mechanistic pairing is not intuitive. But it is precisely that counterintuitive logic — attacking a transcription factor through its downstream DNA interactions rather than the protein directly — that gave this trial its hypothesis. And the phase 1/2 data validates the mechanistic reasoning with clinical signal.

Old assumption: transcription factor fusions are regulatory targets, not drug targets. New assumption: the DNA-binding program of a fusion protein is a pharmacologically accessible vulnerability, and the trial design must be built to capture it.

What the Protocol Architecture Reveals

Running a phase 1/2 combination trial in a rare pediatric and young adult cancer under this mechanistic premise created specific operational and regulatory pressures that sponsors in adjacent rare oncology programs should map carefully.

First, patient stratification here is binary in the best possible way. EWS::FLI1 is present or it is not. There is no ambiguous biomarker expression threshold, no continuous variable requiring cut-point validation, no companion diagnostic negotiation about analytical sensitivity. The fusion is detected by RT-PCR or FISH with established methodology, and it is present in the overwhelming majority of Ewing sarcoma diagnoses. That simplicity is an operational gift in a rare disease setting where enrollment windows are narrow and every screening failure compounds timeline risk.

Compare this to the companion diagnostic burden carried by, for example, a HER2-low breast cancer trial, where the distinction between HER2-low and HER2-zero required years of assay standardization work before Daiichi Sankyo and AstraZeneca could run the DESTINY-Breast04 trial that supported trastuzumab deruxtecan’s label expansion in 2022. The EWS::FLI1 fusion does not create that problem. What it does create — and what this trial had to solve — is the challenge of accruing a statistically meaningful safety and efficacy dataset in a disease with roughly 200 to 300 new diagnoses per year in the United States alone.

That accrual challenge is where the phase 1/2 seamless adaptive structure earns its value. Combining dose-finding and preliminary efficacy evaluation in a single protocol, with pre-specified rules for expansion once the recommended phase 2 dose is identified, reduces the time gap between safety readout and efficacy signal generation. In rare cancers, that gap is not a statistical nuisance — it is a patient survival issue. FDA’s guidance on seamless adaptive trial designs, finalized in 2019, provides the regulatory scaffolding for this kind of combined-phase approach, and this trial’s architecture is a practical case study in applying it to a molecularly defined rare tumor.

Precision Design, Systemic Implications

There is a harder question embedded in this trial’s success that the precision oncology field has not fully answered: what happens when the target is this clean but the disease is this rare?

FDA’s Project Facilitate and the Rare Pediatric Disease priority review voucher program exist precisely because the agency understands that standard review timelines and trial size expectations are calibrated for common diseases. A phase 1/2 trial in Ewing sarcoma will not generate the 400-patient randomized dataset that an FDA reviewer would prefer. It will generate the dataset that the disease population permits. The regulatory question is whether the mechanistic specificity of EWS::FLI1 targeting — a single fusion, a defined transcriptional program, a combinatorial mechanism backed by preclinical rationale — can substitute for statistical power that the patient population structurally cannot provide.

FDA’s guidance on rare pediatric disease drug development acknowledges this tradeoff explicitly, accepting that “substantial evidence” in rare cancers may derive from smaller, mechanistically grounded datasets when the biology is well-characterized. The trabectedin-irinotecan trial is a test of how far that principle extends when the target is a transcription factor fusion rather than a kinase or receptor.

If the efficacy signal in the phase 2 expansion cohort holds at the magnitude suggested by the phase 1 safety and preliminary activity data, this trial becomes a reference case for NDA or BLA construction under accelerated approval pathways — specifically the kind of single-arm, biomarker-selected submission architecture that FDA has accepted in other molecularly defined rare tumors, including the approvals granted to larotrectinib and entrectinib in NTRK fusion-positive cancers.

Sponsors running fusion-driven rare tumor programs — RET fusions in thyroid cancer, FGFR fusions in cholangiocarcinoma, SS18::SSX fusions in synovial sarcoma — should be reading this trial’s methodology section as operational precedent, not just its results.

The forward look is this: the next signal to watch is whether FDA issues a formal development program guidance specific to transcription factor fusion targets, analogous to the guidance framework it built for kinase inhibitors in oncology. The trabectedin-irinotecan EWS::FLI1 data is now the most clinically advanced example of a new therapeutic class. If a BLA submission follows, the agency’s response to the efficacy evidence — what it accepts, what it requires, and whether it issues a complete response letter demanding a randomized confirmatory trial in a disease with 200 annual diagnoses — will define the regulatory floor for every transcription factor program behind it.

References

1. “Trabectedin and low-dose irinotecan to target EWS::FLI1 in Ewing sarcoma: a phase 1/2 trial” (2026)
2. FDA Guidance — Adaptive Designs for Clinical Trials of Drugs and Biologics (2019)
3. FDA Guidance — Rare Pediatric Disease Priority Review Vouchers (current program documentation)
4. FDA Guidance — Rare Diseases: Early Drug Development and the Role of Pre-Investigational New Drug Meetings
5. New England Journal of Medicine — DESTINY-Breast04 trial of trastuzumab deruxtecan in HER2-low breast cancer (Daiichi Sankyo/AstraZeneca, Phase 3, 2022)
6. New England Journal of Medicine — Larotrectinib in TRK fusion-positive cancers (Loxo Oncology, Phase 1/2, 2018)