1,610 patients. 103 tumor types. 37 drugs none of them were approved to receive. A third of them benefited anyway.

Those are the headline results from the Drug Rediscovery Protocol, or DRUP — a decade-long trial across the Netherlands that asked a deceptively simple question: if a cancer drug targets a specific genetic mutation, does it matter which organ the tumor started in?

The answer, published this week in Nature, is a qualified yes — and the implications could reshape how doctors treat patients who have run out of standard options.

What ‘Genomics-Guided’ Actually Means

Most cancer drugs are approved for specific cancer types — lung, breast, melanoma. But cancer is increasingly understood as a disease of genetic mutations rather than anatomical origin. A BRAF mutation in a thyroid tumor is, in molecular terms, the same BRAF mutation found in melanoma. So the logic goes: why not use the melanoma drug for the thyroid cancer patient?

That is genomics-guided prescribing: sequence a patient’s tumor, identify the mutations, and match them to existing drugs — regardless of what the drug’s label says. It is an elegant idea. Without systematic evidence, it is also a gamble.

The Numbers Behind the Trial

DRUP enrolled patients between July 2016 and May 2024 — all with advanced or metastatic solid tumors and no remaining standard treatment. The trial tested 37 approved drugs across 313 cohorts, each defined by a specific drug–target–tumor combination. Among 1,363 evaluable patients, 34.9% showed clinical benefit — either tumor shrinkage or stable disease lasting at least 16 weeks. The objective response rate was 15.7%.

Median overall survival was 8.2 months. For patients who had exhausted all standard options, that is not trivial.

The results varied widely. Of 15 cohorts that completed the trial’s second stage, 14 met pre-defined success criteria. Patients with MET-mutated lung cancer treated with crizotinib saw a 71% clinical benefit rate and 62% objective responses, according to an earlier analysis published by the DRUP investigators. The nivolumab cohort for microsatellite unstable (MSI-H) tumors achieved a 63% benefit rate — strong enough to secure full national reimbursement in the Netherlands for that indication, regardless of tumor type.

Other combinations flopped. CDK4/6 inhibitors palbociclib and ribociclib produced a clinical benefit rate of just 15% and zero objective responses. Those cohorts were closed.

Why Off-Label Prescribing Is Controversial

Off-label use in oncology is already common. An analysis of nearly 166,000 US cancer patients found 18.6% received at least one off-label therapy, according to a recent study cited in the DRUP paper. But it typically happens without systematic data collection, driven more by physician instinct and patient desperation than by evidence.

“Oncologists often treat patients with off-label treatments outside trial settings,” said Emile Voest, principal investigator at the Netherlands Cancer Institute. “This involves significant risks of toxicity, costs and uneven access.”

Grade 3 or higher treatment-related adverse events occurred in 28.4% of DRUP patients. Without the trial’s infrastructure, those harms — and the absence of benefit in roughly two-thirds of participants — would have gone undocumented.

The trial also identified 67 patients — 7% of those with sufficient follow-up — as exceptional responders, achieving complete remission or at least two years of progression-free survival. These outliers, concentrated in MSI-H, BRAF V600E, and high tumor mutational burden subgroups, are the kind of dramatic outcomes the current approval system is poorly equipped to capture.

A System Not Built for This

Drug approval is organized around cancer type. Expanding a drug’s label to other cancers requires new trials — expensive, years-long endeavors with little commercial incentive when patient populations are small.

DRUP offers a middle path: structured evidence from off-label use without the scale of traditional drug development. Its stepwise design — small cohorts that expand only when early results justify it — minimizes exposure to ineffective treatments.

The model is spreading. Nordic countries have adopted DRUP-like protocols, and two EU-funded initiatives, PCM4EU and PRIME-ROSE, aim to expand the approach across Europe.

The Hard Caveat

The trial was not randomized. There was no control group — deliberately, since withholding potentially effective drugs from patients with no other options raises clear ethical problems. But the absence of a comparator makes it harder to distinguish drug effects from the natural course of disease.

Two-thirds of patients derived no clinical benefit. Some experienced serious side effects for their trouble.

The investigators’ recommendation is direct: off-label anticancer drugs should only be prescribed within frameworks that systematically evaluate outcomes. The evidence that genomics-guided treatment works — for some patients, some of the time — is now considerably stronger. The hard work of figuring out exactly which patients, and when, is just beginning.

Sources