The largest off-label cancer drug study found modest benefit, real for a minority

Doctors prescribe cancer drugs off-label constantly — for a tumor genetically similar to the one a drug was approved for, when a patient has run out of standard options and a genomic test suggests a plausible target. What has been missing, largely, is systematic evidence of whether that practice actually works. A study published April 15 in Nature is the largest attempt yet to answer that question directly, tracking outcomes for 1,610 patients across 37 different off-label drugs [s1].

The headline finding is genuinely two findings, and they point in different directions depending on which patient you're describing.

What the Drug Rediscovery Protocol did

The Drug Rediscovery Protocol, known as DRUP, is a Dutch trial that has prospectively enrolled patients with advanced solid tumors who have exhausted standard treatment options and carry a genomic alteration believed to make them plausible candidates for a drug approved for a different cancer [s1]. Between July 2016 and May 2024, 1,610 patients began treatment through the trial; 1,363 were response-evaluable, including 533 (39.1%) with cancers rare enough to make running a dedicated trial for them impractical [s1].

Unlike a conventional randomized trial testing one drug against a comparator, DRUP is a registry-style framework: each patient is matched to one of 37 different off-label drugs based on their tumor's genomic profile, and outcomes are tracked and pooled across all of them [s1].

The topline numbers

Across the full evaluable population, the clinical benefit rate — confirmed response or stable disease lasting at least 16 weeks — was 34.9% (95% CI, 32.2-37.6), and the objective response rate was 15.7% (95% CI, 13.7-17.9) [s1]. Median progression-free survival was 3.4 months; median overall survival was 8.2 months [s1]. Grade 3 or higher treatment-related adverse events occurred in 28.4% of patients [s1].

Those are the numbers for a population that, by definition, had already exhausted standard treatment options — so they should be read against that baseline rather than against outcomes in newly diagnosed cancer. Even so, the authors' own characterization is blunt: activity across all tumor-drug combinations pooled together was "modest" [s1].

The subgroup that did well

Modest on average does not mean modest for everyone. The paper reports that 7.0% of evaluable patients were "exceptional responders" — a defined subgroup within specific molecular subgroups that achieved meaningful, sustained benefit far beyond the trial's average [s1]. The paper frames this as the central practical lesson: outcomes varied enormously by which genomic alteration and which drug were matched, and some of those matches worked dramatically better than others, while many did not work at all.

That variation is also the paper's argument for why off-label precision oncology needs structure rather than ad hoc prescribing. A clinical benefit rate of 34.9% pooled across 37 drugs could mean every combination works modestly, or it could mean — as DRUP's data suggest — that a handful of combinations work very well and most do not work much better than would be expected by chance. Only systematic tracking distinguishes those two very different underlying realities, which is exactly what most off-label prescribing outside a structured trial does not generate.

Real-world consequences already in motion

This is not purely an academic exercise. The authors report that evidence generated within DRUP has already been used by Dutch regulatory bodies to inform reimbursement decisions for off-label cancer drug use [s1] — meaning the trial's findings are functioning as a substitute for the kind of evidence base that individual off-label prescribing decisions don't typically generate on their own.

What the authors recommend

The paper's authors do not conclude that off-label precision prescribing should stop, nor that it should be treated as equivalent to on-label, trial-proven use. Instead, they recommend that off-label precision medicines be used only within frameworks that systematically evaluate both efficacy and toxicity, support refinement of which biomarkers actually predict benefit, and enable a stepwise path toward formal label expansion where evidence supports it — prioritizing high-confidence genomic targets, early intervention, endpoints aligned with regulatory standards, and international collaboration to pool enough patients for rare tumor-drug combinations to be evaluated at all [s1].

The limits of what this study shows

DRUP is not a randomized trial, and it does not include a comparator arm — there is no way, from this data alone, to know how these patients would have fared on best supportive care or a different off-label choice. The 39.1% of participants with rare cancers is itself a reminder of why: for many of these tumor-genotype combinations, a conventional randomized trial with adequate statistical power may never be feasible. That is the gap DRUP's registry approach is built to partially fill, not a substitute for definitive trial evidence where such trials are possible.

This article is informational and is not medical advice.

Sources

• [s1] Verkerk K, Spiekman AC, Haj Mohammad SF, et al. (DRUP Trial Investigators). "Prospective evaluation of genomics-guided off-label treatment." Nature, 15 April 2026. https://doi.org/10.1038/s41586-026-10405-x