Sample benchmark sheet · Oncology
Refractory metastatic colorectal cancer: OS in Phase 3 trials
What published trials showed in this setting. This is precedent, not an FDA requirement.
- Indication
- Colorectal Cancer
- Endpoint
- Overall survival (OS)
- Setting
- Refractory, after prior standard therapy
- Checked
- 2 November 2026
Lowest approved precedent
+1.4 months
Gain in median OS over the control arm
CORRECT: 6.4 vs 5.0 months
The weakest result among the approved trials on this sheet. Approved precedents listed here showed median OS gains of +1.4 months (CORRECT), +1.8 months (RECOURSE) and +3.3 months (SUNLIGHT, over trifluridine/tipiracil alone). All three enrolled refractory disease after prior standard therapy. Observed precedent, not an FDA bar.
The trials
Each of these trials supported an FDA approval.
| Trial | Result | Sources |
|---|---|---|
| RECOURSEtrifluridine/tipiracil vs placebo, refractory | 7.1 vs 5.3mFigures from: Paper abstract | Mayer et al., NEJM 2015 |
| SUNLIGHTtrifluridine/tipiracil+bevacizumab vs trifluridine/tipiracil, refractory, up to 2 prior regimens | 10.8 vs 7.5mFigures from: Paper abstract | Prager et al., NEJM 2023 |
| CORRECTregorafenib vs placebo, after standard therapies | 6.4 vs 5.0mFigures from: Paper abstract | Grothey et al., Lancet 2013 |
Results are median OS, experimental arm vs control arm. m: months.
Observed range: 6.4 months to 10.8 months, median 7.1 months. Based on the 3 named trials listed. Range and median use their experimental arms. First-line and MSI-H trials are excluded.
The limits of this sheet
- Control arms differ. CORRECT and RECOURSE compared with placebo. SUNLIGHT compared with trifluridine/tipiracil, an active drug, so its 3.3 months is on top of that drug.
- A gain in median months is not the whole effect. Hazard ratios were 0.77 in CORRECT, 0.68 in RECOURSE and 0.61 in SUNLIGHT.
- Lines of therapy differ. CORRECT enrolled patients after all standard therapies, while SUNLIGHT allowed no more than two prior chemotherapy regimens.
- First-line and MSI-H trials are left out. So is fruquintinib's FRESCO-2, which is on a separate row in the app.
- This is precedent, not a requirement. These results supported approvals in their own context, and matching a number here guarantees nothing.
- This is a selected list. Only the trials named here are counted, and other approvals in this setting are not on this sheet.
- Results across trials are not directly comparable. Populations, control arms, follow-up and response criteria differ.
FDA context
OS was the primary endpoint in CORRECT, RECOURSE and SUNLIGHT, and gains of +1.4 to +3.3 months supported approval in refractory disease. Note: KEYNOTE-177 (pembrolizumab 1L MSI-H CRC) is excluded because it is a first-line biomarker-selected trial, and its OS (NR vs 36.7m) did not reach the prespecified significance level.
Also cited on this row
How this was checked
On 2 November 2026 automated tools checked each result against the source named beside it and resolved each paper’s DOI. No clinician has reviewed this sheet. Check the source before you rely on a figure. Tell us if something looks wrong.
Look up an indication you know well
This is a sample. The app has the rest of our benchmark library, which is mostly oncology, and tracks changes to the ClinicalTrials.gov records of the programs you watch. It is free during early access, with no card needed.