Patients with advanced Ewing sarcoma or osteosarcoma survive less than a year on average, and no standard salvage strategy exists. The single-arm phase 2 CABONE trial tested cabozantinib — an oral MET/VEGFR2 inhibitor — in 90 heavily pre-treated patients across ten French Sarcoma Group centers. The drug drove partial responses in roughly a quarter of Ewing sarcoma patients and kept a third of osteosarcoma patients progression-free at six months, with a manageable safety profile. Exploratory plasma biomarker analysis nominated baseline VEGF-A and soluble MET as candidate predictors of outcome, offering a way to enrich future trials.
The CABONE study — sponsored by Institut Bergonié (Bordeaux) and led by Prof. Antoine Italiano — was run by the French Sarcoma Group across ten centers, with cabozantinib supplied by the US National Cancer Institute’s Cancer Therapy Evaluation Program (CTEP) under a research agreement with Exelixis. It was funded by the French National Cancer Institute (INCa) and l’Association pour la Recherche contre le Cancer. Alongside the clinical readout, the trial included a prespecified, exploratory analysis of circulating biomarkers — VEGF-A, hepatocyte growth factor (HGF), soluble VEGFR2, and soluble MET — measured in patient plasma.
CABONE enrolled 90 patients (45 Ewing sarcoma, 45 osteosarcoma), aged 12 and up, with documented disease progression and largely refractory disease (median 2 prior lines). All received oral cabozantinib (60 mg daily; 40 mg/m² for patients under 16) in 28-day cycles until progression or unacceptable toxicity, with response assessed by blinded central RECIST 1.1 review. 39 Ewing sarcoma and 42 osteosarcoma patients were efficacy-evaluable.
Ten of 39 Ewing sarcoma patients (26%; 95% CI 13–42) achieved an objective response by 6 months — all partial responses — meeting the primary endpoint. A further 49% had stable disease, with tumor shrinkage seen in 38% overall. Median progression-free survival was 4.4 months and median overall survival 10.2 months in this heavily pre-treated group.
In osteosarcoma, five of 42 patients (12%; 4–26) had a partial response and 14 (33%; 20–50) were progression-free at 6 months, satisfying the dual primary endpoint. Four-month progression-free survival reached 71% — above the 30% activity threshold benchmarked for advanced osteosarcoma agents. Pneumothorax occurred in 13% of the full cohort (all with pleural or subpleural metastases at baseline) and was generally manageable.
Explicyte’s contribution centered on the prespecified plasma biomarker analysis of VEGF-A, HGF, soluble VEGFR2, and soluble MET (sMET), evaluable in 35 Ewing sarcoma and 36 osteosarcoma patients. In osteosarcoma, low baseline VEGF-A (<12.5 pg/mL) was associated with longer overall survival (13.2 vs 8.2 months; log-rank P = 0.014), and baseline sMET (cutoff 300.6 ng/mL) stratified progression-free survival (7.8 vs 5.4 months; log-rank P = 0.016). No biomarker was associated with outcome in the Ewing sarcoma cohort.
Cabozantinib showed a genuine activity signal in two rare, refractory sarcomas that lack standard salvage therapy — and the trial's plasma biomarker work points to a route for selecting patients in future studies.
For drug developers, CABONE establishes cabozantinib as an active, tolerable option worth further investigation in advanced Ewing sarcoma and osteosarcoma — diseases where randomized trials are hard to power because of their rarity. Just as important for translational teams, the exploratory circulating-biomarker analysis shows that baseline VEGF-A and soluble MET may help identify likely responders, arguing for prospective plasma biomarker collection in the next generation of sarcoma trials. Enriching for biomarker-defined subsets could sharpen signal detection where objective response alone is an imperfect readout.
