Daraxonrasib Challenges Second-Line Pancreatic Cancer Care

It still needs a comparison strong enough to support a change in routine care.

Start with the two numbers

The daraxonrasib result comes from a peer-reviewed comparative report indexed in PubMed. Daraxonrasib, also known as RMC-6236, is an investigational pan-RAS inhibitor. Among previously treated patients with metastatic pancreatic cancer harboring a KRAS mutation, reported median overall survival was 13.2 months with the drug and 6.6 months with chemotherapy.

I wrote both figures down because a 6.6-month absolute difference is hard to dismiss in a setting where survival after first-line treatment is often measured in months. The reported medians have a ratio of 2.0. That does not mean daraxonrasib doubled the life of each person who received it.

Median survival is the point at which half the patients in a group have died. Some patients live longer, some less, and the median cannot tell a person what will happen to them. It also cannot, by itself, show that treatment caused the difference between groups.

That second limitation is the one I kept circling in the notebook. A large contrast can remain clinically interesting even when the comparison is uncertain, but the uncertainty matters more, not less, when the result is large enough to influence treatment conversations.

A ratio of medians is descriptive. It is not a substitute for a hazard ratio, which uses the timing and pattern of events across follow-up rather than reducing each survival curve to one point. The full Kaplan-Meier curves, numbers at risk and tests of the proportional-hazards assumption would help show whether the apparent advantage persisted or came mainly during one portion of follow-up.

The comparison carries the weight

The relevant population is adults receiving second-line treatment for KRAS-mutant metastatic pancreatic cancer. Daraxonrasib is a RAS(ON) multi-selective inhibitor designed to act against activated RAS signaling rather than one KRAS variant alone, and its development program has included people with advanced solid tumors carrying selected RAS alterations.

The comparative report identifies chemotherapy as the control. From the citation-level information available for this brief, I could not verify the analytic sample size, the chemotherapy regimens, median follow-up, a hazard ratio, confidence intervals or a p value. These are not stray details. A small cohort can produce an unstable median, and limited follow-up can leave a survival estimate immature, while a broad label such as chemotherapy may combine treatments that do not perform the same way.

The first question for the full report is whether patients were assigned concurrently and at random. If the daraxonrasib group came from a prospective single-arm trial while the chemotherapy group came from historical records, routine-care data or another external source, then the analysis was nonrandomized even if investigators used matching, weighting or statistical modeling.

Those methods can reduce measured differences. They cannot repair information that was never collected, recorded differently or modeled poorly.

Performance status is particularly important in metastatic pancreatic cancer. Disease burden and liver metastases may matter, as can prior treatment, the response to that treatment and the time between diagnosis and second-line therapy. KRAS variant, laboratory status, access to specialist care and the ability to enter a clinical trial may also separate the groups in ways that affect survival apart from the drug itself.

Patients entering an early-phase study have generally survived long enough and remained well enough to enroll. They may not resemble every patient who receives second-line chemotherapy in regular practice. I added that point beside the two survival figures in the notebook, then left the question of how much it changed the result unanswered. The available citation information cannot settle it.

Why 13.2 months stands out

Randomized trials of established chemotherapy provide context, although they are not direct controls for daraxonrasib.

In NAPOLI-1, liposomal irinotecan combined with fluorouracil and leucovorin produced median overall survival of 6.1 months, compared with 4.2 months for fluorouracil and leucovorin after gemcitabine-based treatment. The reported hazard ratio was 0.67, with a 95% confidence interval from 0.49 to 0.92.

CONKO-003 reported median overall survival of 5.9 months with oxaliplatin, folinic acid and fluorouracil, versus 3.3 months with folinic acid and fluorouracil in gemcitabine-refractory disease. Its hazard ratio was 0.66, with a 95% confidence interval from 0.48 to 0.91.

These trials make a chemotherapy median near six months plausible. They do not prove that the chemotherapy patients in the daraxonrasib comparison were equivalent to those randomized populations, nor can separate trials establish how much better one treatment would have been for the same patients at the same point in illness.

Still, 13.2 months exceeds those historical randomized benchmarks by a meaningful margin. The result supports continued development and a definitive head-to-head trial. It may also support a discussion of clinical-trial enrollment when a suitable study is available.

It does not establish daraxonrasib as the replacement for standard second-line chemotherapy across all KRAS-mutant pancreatic cancers. Treatment choices depend on prior therapy and performance status, along with comorbidities, molecular findings, patient preferences, access and regulatory status. Publication of a comparative analysis does not create an FDA-approved indication.

Survival is not the only comparison

The notebook line is about survival, but patients live through treatment, not through a median.

Serious adverse events, treatment interruptions and discontinuations need to be read beside the survival curves. Symptom control and quality of life also matter, particularly in metastatic pancreatic cancer, where disease symptoms and treatment burden can overlap. Pan-RAS inhibition may feel different from cytotoxic chemotherapy even if the percentages of severe adverse events appear similar on a table.

Comparative patient-reported outcomes would help. So would a clear account of why patients stopped treatment and whether follow-up captured those experiences in the same way for both groups.

Assessment schedules may differ between a clinical trial and an external chemotherapy cohort. Trial participants often undergo planned imaging and regular follow-up, while routine-care records may document progression, discontinuation or death less consistently. Overall survival is less subjective than response rate or progression-free survival, but missing death information, unequal follow-up and different starting points can still shift an estimate.

What a confirming trial needs to show

The strongest confirmation would be an adequately powered randomized trial comparing daraxonrasib with a specified, clinically appropriate second-line regimen. Randomization would need to account for major prognostic differences, including first-line treatment and performance status, while also addressing relevant features of the disease.

Overall survival should remain central. Progression-free survival, response and duration of response would add context, while quality-of-life findings and safety results would show what patients experienced during the additional time. The analyses should be prespecified rather than selected after investigators see where the apparent differences fall.

Molecular generalizability needs its own attention. The term KRAS-mutant covers multiple variants and biologic settings. Investigators should report outcomes by variant without treating small subgroups as settled evidence, and they should examine whether co-alterations or resistance mechanisms are associated with greater or lesser benefit.

Trial eligibility is another boundary. Results from people fit enough to enroll may not transfer fully to patients with poor performance status, organ dysfunction or extensive comorbidity. That problem will not disappear even after a positive randomized trial, though broader enrollment and transparent reporting can make the limits easier to see.

The maturity of the survival data also has to be visible. Readers need the number of deaths, length of follow-up, censoring pattern and numbers at risk, together with confidence intervals around each median. Independent replication would strengthen confidence. Funding, investigator relationships and major analytic decisions should be reported plainly.

Questions clinicians ask

Should daraxonrasib replace second-line chemotherapy now?

Not on the two median survival figures alone. The result is strong enough to justify further comparative testing and consideration of trial participation, but a practice change requires confidence that the groups were comparable, the estimate was statistically precise and the findings apply to the patient in front of the clinician. Safety, quality of life and regulatory status remain part of that judgment.

How should I explain the 13.2-month median to patients?

Half of the daraxonrasib group was alive at 13.2 months in the reported analysis. The figure does not predict an individual outcome, and it does not show that every person gained 6.6 months. How much benefit the drug caused depends on whether the treatment groups were genuinely comparable.

Does the finding apply to every KRAS mutation?

That cannot be assumed from a pooled KRAS-mutant result. Variant-level enrollment and survival findings are needed, along with information about prior treatment and relevant co-alterations. Estimates from small molecular subgroups may move substantially as more events are recorded.

What evidence would be most likely to change practice?

A randomized head-to-head trial showing a statistically precise overall-survival benefit would carry the most weight, particularly if toxicity was acceptable and quality of life was preserved. Clinicians would also need enough follow-up to see whether the difference lasted and whether patients across common prior-treatment pathways were represented.

My notebook remains open to the same two lines: 13.2 months, then 6.6 months.

Questions people ask

Did daraxonrasib double survival compared with chemotherapy?

The reported median survival times were 13.2 months with daraxonrasib and 6.6 months with chemotherapy, giving a ratio of 2.0. That does not mean every patient lived twice as long, and the medians alone cannot show that the drug caused the difference.

Should daraxonrasib replace second-line chemotherapy now?

The author concluded that the two median survival figures are not enough to support replacement of standard chemotherapy. Daraxonrasib remains investigational, and key details about the comparison, statistical precision, safety and follow-up were not verified from the available citation information.

Does the daraxonrasib result apply to every KRAS mutation?

The story says that cannot be assumed from a pooled KRAS-mutant result. Variant-level outcomes, prior-treatment details and relevant co-alterations would be needed, while small molecular subgroup estimates may change as more events are recorded.

References

1. Daraxonrasib (RMC-6236) doubles survival vs chemotherapy in second-line metastatic pancreatic cancer — PubMed, 2026
2. Study of RMC-6236 in Patients With Advanced Solid Tumors Harboring Specific Mutations in RAS — ClinicalTrials.gov, 2022
3. Nanoliposomal irinotecan with fluorouracil and folinic acid in metastatic pancreatic cancer after previous gemcitabine-based therapy (NAPOLI-1): a global, randomised, open-label, phase 3 trial — The Lancet, 2016
4. Second-line oxaliplatin, folinic acid, and fluorouracil versus folinic acid and fluorouracil alone for gemcitabine-refractory pancreatic cancer: outcomes from the CONKO-003 trial — Journal of Clinical Oncology, 2014
5. Pancreatic Cancer Treatment (PDQ)—Health Professional Version — National Cancer Institute, 2026