Orbus Bet $71 Million on One Brain-Cancer Trial - Then the Definition of the Disease Changed

Biotech likes a clean result. A drug works or it does not. A trial wins or it loses. Orbus Therapeutics landed somewhere far more interesting and much less comfortable. Its only late-stage drug candidate, eflornithine, went through a 343-patient Phase 3 trial for recurrent anaplastic astrocytoma. In the whole study population, adding the drug to lomustine did not significantly extend survival. That is the first fact. The second is that in patients whose tumors fit a newer molecular definition - IDH-mutant, grade 3 astrocytoma - median survival was 34.9 months with the combination and 23.5 months with lomustine alone.

Both facts are true. Neither lets Orbus skip the hard work ahead. Together, they explain why this unusually small Palo Alto company has survived long enough to become a case study in what happens when science redraws a market while a startup is still testing its product.

One molecule, one company-sized wager

Orbus was founded in 2012 by Bob Myers and Jason Levin, two operators with roots in Jazz Pharmaceuticals and ALZA. The original pitch was narrow: develop products for rare, serious diseases with few effective therapies. Its operating reality became narrower still. The company centered itself on eflornithine, an inhibitor of ornithine decarboxylase, an enzyme involved in making polyamines that cells use for growth and survival.

The molecule was not new. Injectable eflornithine had been approved for African sleeping sickness. A cream was approved to slow unwanted facial-hair growth. In 2023, another oral formulation was approved in the United States to reduce relapse risk in high-risk neuroblastoma. Orbus’ idea was different: use an oral solution with lomustine to slow the growth of certain recurrent brain tumors.

That distinction matters commercially. Orbus is not selling software seats or a service contract. It is precommercial. Its business is to manufacture evidence, intellectual property and regulatory credibility around a known compound in a specific use. It has patents covering oral formulations and has pursued methods of treating gliomas. If approved, it could commercialize the regimen itself or license it to a larger drugmaker. Until then, patients are trial participants, physicians are investigators and revenue is an aspiration.

What failed first

The Phase 3 study, named STELLAR, began in 2016. Adults whose anaplastic astrocytoma had returned after radiation and temozolomide were randomized to receive lomustine alone or lomustine plus oral eflornithine. The study crossed an interim futility review in 2021, enrolled 343 people across 74 sites in eight countries and completed follow-up. This was not a boutique signal-hunting exercise. It was the company’s pivotal test.

The first thing to fail was the headline hypothesis. In the intention-to-treat population, median overall survival was 23.4 months with the combination and 20.3 months with lomustine. The hazard ratio was 0.94 and the difference was not statistically significant. Progression-free survival was also not different across the broad group. A disciplined reading stops here long enough to say plainly: STELLAR missed its primary endpoint.

The broad result failed. The useful question became: was the population too broad?

The combination also carried a price that is not measured in dollars. Grade 3 or worse events involving reversible bone-marrow suppression occurred in 42 percent of combination patients versus 29 percent on lomustine alone. Hearing impairment of that severity appeared in 24 percent of the combination group and none of the control group. Any future use must earn those risks with a convincing benefit in the right patients.

The diagnosis changed underneath the trial

When STELLAR started, anaplastic astrocytoma was still substantially defined by what pathologists saw under a microscope. During the trial, brain-tumor classification shifted toward molecular identity. The World Health Organization’s 2021 framework made mutations in the IDH gene central to naming adult diffuse gliomas and reclassified some tumors that would previously have shared a diagnosis.

Orbus did not change the molecule. The field changed the map. Before the study was unblinded, investigators defined a subgroup using the newer criteria. In the peer-reviewed analysis, 196 patients had recurrent IDH-mutant grade 3 astrocytoma. Median overall survival was 34.9 months with eflornithine plus lomustine versus 23.5 months with lomustine. Median progression-free survival was 15.8 months versus 7.2 months. No comparable benefit appeared in grade 4 disease.

This was not a random after-the-fact slice such as “left-handed patients enrolled on Tuesdays.” IDH status is a foundational part of modern tumor biology, and the subgroup was defined before unblinding. Even so, it remains a subgroup after a negative overall result. Regulators will decide what additional evidence is needed. Doctors will weigh survival, uncertainty and toxicity. Orbus cannot declare the answer on their behalf.

Where Orbus fits - and what it costs

Recurrent grade 3 astrocytoma is a small, specialist market with limited standard options. Care may involve repeat surgery, radiation, alkylating chemotherapy and clinical trials, depending on the tumor and the patient. Lomustine is the practical comparator, not a flashy venture-backed rival. Orbus is differentiated by its metabolic target, oral formulation and Phase 3 dataset in a molecularly selected group.

The financial cost is easier to count than the strategic one. Orbus raised an initial $32.5 million Series A in 2015. By 2020, the expanded round totaled $71 million, backed by Longitude Capital, H.I.G. BioVentures, Adams Street Partners and Abingworth. The company put that capital behind a global trial, manufacturing and an intellectual-property portfolio. It also accepted concentration risk: one lead product means one clinical result can redraw the value of the entire company.

Its compact team is part of the model. Public leadership materials show specialists in regulatory affairs, product development and oncology, with consultants filling senior clinical roles. The expertise is less about discovering thousands of compounds than shepherding one old molecule through formulation, multicountry operations, biomarker analysis and regulatory argument. Customers, if approval ever arrives, would be neuro-oncologists and the relatively small number of patients whose recurrent tumors match the molecular label.

That makes the addressable market small by consumer standards and consequential by orphan-drug standards. Orbus does not need a mass audience; it needs diagnostic testing to identify the right tumor, specialist centers able to manage a demanding regimen and payers willing to reimburse a treatment backed by persuasive evidence. The therapy would also enter care after surgery, radiation and temozolomide, not replace every earlier step. Its place is a particular fork in a long clinical road. The sharper that fork becomes, the more legible Orbus’ value proposition is - and the less room the company has for loose claims about whom the drug helps.

The lesson has boundary conditions

The Orbus playbook does not work just because an old drug is cheap or a post-hoc chart looks attractive. Repurposing fails when the biological rationale is weak, when formulation cannot create reliable exposure, when patents cannot support investment or when the target population is too small to recruit and serve. A subgroup strategy fails when the marker was discovered by rummaging through noise, when there is no mechanistic reason for the split or when toxicity overwhelms the possible gain.

It also fails if a company confuses designation with approval. Orphan and Breakthrough Therapy status can bring incentives and closer regulatory interaction; neither proves efficacy or puts medicine on a pharmacy shelf. Orbus has a peer-reviewed, randomized data set and a coherent molecular hypothesis. It does not publicly have an FDA approval for eflornithine in astrocytoma.

What changed the company’s mind was not a sudden pivot away from its drug. It was better disease taxonomy. The broad label “anaplastic astrocytoma” had mixed tumors that modern medicine understands as biologically different. Once the categories became sharper, so did the apparent treatment effect. This is a useful idea far beyond oncology: if your customer group is defined by appearances rather than causes, the average may hide the people for whom your product matters.

Orbus now occupies biotech’s least cinematic middle chapter. It has moved beyond a clever theory but has not reached an approved product. The company has evidence worth taking seriously, an overall miss that cannot be edited away and a narrower population that may justify the next bet. That is less tidy than a victory. It is also closer to how scientific progress usually looks.