Daraxonrasib or Chemotherapy in Previously Treated Metastatic Pancreatic Cancer
Peer-reviewed phase 3 trial; NEJM 2026, 2026 · paper · Clinical trials
A phase 3 randomized trial of oral RAS inhibitor daraxonrasib versus chemotherapy in 500 patients with previously treated metastatic pancreatic cancer. Median overall survival was 13.2 versus 6.6 months in the RAS G12 population (hazard ratio 0.40).
1 finding
Wording The Discussion says "the percentage of patients with a response was twice as high with daraxonrasib", but the Results give 33.2% vs 11.8% (RAS G12) and 31.6% vs 11.2% (overall), about 2.8-fold. Also, none of the four response percentages can be obtained from a whole-number count over N=228/231/248/252.
- p. 5 Results, Objective Response
In the RAS G12 population, the objective response was 33.2% (95% CI, 27.0 to 39.9) with daraxonrasib
- p. 7 Results, Objective Response (continued)
as compared with 11.8% (95% CI, 7.8 to 16.8) with chemotherapy (Table S2). In the overall population, the response was 31.6%
- p. 11 Discussion, first paragraph
percentage of patients with a response was twice as high with daraxonrasib
- p. 7 Results, Objective Response
as compared with 11.8% (95% CI, 7.8 to 16.8) with chemotherapy
Only the authors can say what is right here.
Judge's note: Checked p. 5, p. 7 and the p. 11 crops. In both populations the ratio is about 2.8, so "twice as high" understates the difference as printed. The response percentages were also checked against the population sizes in Table 1 (228, 231, 248, 252), and the Figure 1 legend says "The efficacy analyses were based on data from all the patients who had undergone randomization." No whole-number count reproduces 33.2%, 11.8%, 31.6% or 11.2% over those denominators. The main text defines no response-evaluable subset. Table S2, which is cited for these figures, is not among the files, so only the authors can say which denominator was used.
Second check: another model agreed
Report a problem with this finding
5 minor (typos and the like; most checked by the judge only)
Wording For chemotherapy, Table 2 gives "Treatment-related adverse event leading to treatment discontinuation ... 24 (11.2)" (also in the abstract and text), but Figure 1 lists only "14 (5.6%) Had adverse event" among the reasons for discontinuing treatment.
- p. 10 Table 2, 'Treatment-related adverse event leading to treatment discontinuation' row, chemotherapy column
24 (11.2)
- p. 5 Figure 1, chemotherapy discontinuation box
14 (5.6%) Had adverse event
- p. 1 Abstract, Results
in 11.2% of those in the chemotherapy group
Only the authors can say what is right here.
Judge's note: Checked the Table 2 row (p. 10) and the Figure 1 flow box (p. 5). Discontinuations for treatment-related adverse events (24) outnumber discontinuations for any adverse event in Figure 1 (14). Figure 1 seems to record one primary reason per patient: 102+29+26+14+5 = 176, plus 3 for other reasons = 179. Table 2 may count stopping any single component of the regimen. However, neither the Table 2 footnotes nor the text says so. In the daraxonrasib group the figures are consistent (3 vs 8). This is a query for the authors.
Second check: another model agreed
Report a problem with this finding
Wording The Discussion says of dermatologic and GI events "among these, events of grade 3 were uncommon", while grade ≥3 rash (13.7%) and stomatitis (12.0%) were the only grade ≥3 treatment-related events reported in at least 10% of daraxonrasib patients.
- p. 12 Discussion, safety paragraph
events of grade 3 were un-common, and no events of grade 4 were report-ed
- p. 10 Table 2, Rash and Stomatitis rows, daraxonrasib
206 (85.5); 33 (13.7)
- p. 8 Results, Safety
Treatment-related adverse events of grade 3 or higher that occurred in at least 10% of the pa-tients in the daraxonrasib group were rash
- p. 12 Discussion, safety paragraph
no events of grade 4 were report-ed
Judge's note: The "no events of grade 4" claim is scoped by "among these" to dermatologic and gastrointestinal events. The grade 4 events in Table 2 footnotes † and ‡ are increased AST and decreased platelet count, which are neither dermatologic nor GI, so that part is consistent and the reviewers' grade 4 contradiction is a misreading. The remaining point is wording: calling grade 3 rash and stomatitis "uncommon" when each exceeds 10% is debatable, but defensible relative to all-grade incidence (85.5% and 53.1%).
Second check: another model agreed
Report a problem with this finding
Typo Methods, End Points (p. 3) read "EORTC Core Quality of Life Questionnaire–Core 30 [EORTC QLQ-C30]". The first "Core" is redundant.
- p. 3 Methods, End Points
EORTC Core Quality of Life Questionnaire–Core 30 [EORTC QLQ-C30]
Judge's note: Checked on the page-3 crop. The abbreviation QLQ-C30 and the parallel instrument name in the same paragraph ("European Organization for Research and Treatment of Cancer Quality of Life Questionnaire pancreatic cancer pain scale") both point to "Quality of Life Questionnaire–Core 30", so the extra "Core" before "Quality" is a duplicated word.
Second check: none (judge only)
Report a problem with this finding
Typo Table 1 footnote ‖ reads "33 patients in chemotherapy group who received", which is missing "the" ("in the chemotherapy group").
- p. 7 Table 1 footnote ‖
33 patients in chemotherapy group who received
Judge's note: The crop of the page-7 footnote shows the article is missing. Earlier in the same sentence the text reads "in the daraxonrasib group".
Second check: none (judge only)
Report a problem with this finding
Typo The Figure 3A panel title reads "Progression-free Survival in the RAS G12 population" with a lowercase "population". Figures 2A, 2B and 3B capitalise "Population".
- p. 9 Figure 3, panel A title
Progression-free Survival in the RAS G12 population
Judge's note: The page-9 crop shows the lowercase "population". The text layer confirms that the other three panel titles use "Population".
Second check: none (judge only)