Skip to content
Bending the Curve Trial by Trial
?

Colon (digestive system)

Colon cancer

Colon cancer care from surgery to immunotherapy; also covers screening, genetics and metastatic disease for colon and rectum together.

Colon and rectum combined · 5-year relative survival · United States, SEER 8 registries

49.41%→67.73%

1975 (observed value) to 2018 (observed value; last observed year in the table). Read these with the caveats under “Then and now”.Source

52 milestones, by decade

  • NordICC randomised presumptively healthy people aged 55 to 64, drawn from population registries in Poland, Norway, Sweden and the Netherlands, to an invitation to one screening colonoscopy or to no invitation. In the three countries with follow-up data (84,585 participants), 11,843 of 28,220 people in the invited group (42.0%) were screened. After a median of 10 years, the risk of colorectal cancer was 0.98% in the invited group and 1.20% in the usual-care group (risk ratio 0.82; 95% CI 0.70 to 0.93) and the risk of death from colorectal cancer was 0.28% versus 0.31% (risk ratio 0.90; 95% CI 0.64 to 1.16). These are intention-to-screen results for an invitation, with fewer than half of invitees screened, and the confidence interval for cancer death (0.64 to 1.16) spans a risk 36% lower to 16% higher.
  • KEYNOTE-177 randomised 307 people with previously untreated MSI-H or dMMR metastatic colorectal cancer to pembrolizumab (153) or investigator-choice chemotherapy (154), with crossover to pembrolizumab allowed after progression. At the second interim analysis, median progression-free survival was 16.5 months with pembrolizumab and 8.2 months with chemotherapy (hazard ratio 0.60; 95% CI 0.45 to 0.80), and grade 3 or higher treatment-related adverse events occurred in 22% and 66% of patients, respectively. The final overall-survival analysis did not meet its prespecified significance threshold (hazard ratio 0.74; 95% CI 0.53 to 1.03), and the report at more than 5 years gives an effective crossover rate of 62% in the chemotherapy group, so most people assigned to chemotherapy also went on to anti-PD-1 or anti-PD-L1 therapy, which complicates the survival comparison.
  • CHALLENGE randomised 889 people with resected colon cancer who had completed adjuvant chemotherapy to a 3-year structured exercise programme (445) or health-education materials alone (444). At a median follow-up of 7.9 years, disease-free survival was longer with exercise (hazard ratio for disease recurrence, new primary cancer or death 0.72; 95% CI 0.55 to 0.94; 5-year disease-free survival 80.3% versus 73.9%). Overall survival was not the primary endpoint: the report describes its result (hazard ratio for death 0.63; 95% CI 0.43 to 0.94) as supporting longer survival and gives no P value for it. Musculoskeletal adverse events were more frequent with exercise (18.5% versus 11.5%).

Evidence current to October 2026 · Checked against primary sources to the extent possible

A guided tour of 12 milestones. You can pause or step through it yourself.

Filter milestonesshowing all 52
Treatment type
Setting

Chemotherapy

  1. 1957 · Treatment · Chemotherapy⚑ needs reviewHeidelberger 1957: fluorinated pyrimidines as tumour inhibitorsIn 1957 Heidelberger and colleagues published a short paper in Nature titled 'Fluorinated pyrimidines, a new class of tumour-inhibitory compounds'. This record rests on the title only, because no abstract or text was read. Fluorouracil is now widely used in cancer treatment, and it is the backbone of the colon cancer adjuvant regimens later in this timeline.Heidelberger C, Chaudhuri NK, Danneberg P et al., Nature 1957, 'Fluorinated pyrimidines, a new class of tumour-inhibitory compounds'. PubMed holds the title and bibliographic record only (no abstract) and the text was not read, so no experimental design or results are recorded. A 2003 review describes 5-fluorouracil as widely used in cancer treatment.

Molecular origins

  1. 1990 · Understanding · Understanding⚑ needs reviewAdenoma-carcinoma sequence and a genetic model of colorectal cancerA 1988 study of 172 bowel tumour samples by Vogelstein and colleagues found that DNA changes accumulated as polyps progressed towards cancer: some were common in large polyps and cancers but rare in small polyps. In 1990 Fearon and Vogelstein published a Cell paper presenting a genetic model of colorectal tumour development.Vogelstein 1988 (172 specimens): ras mutations in 58% of adenomas over 1 cm, 47% of carcinomas and 9% of smaller adenomas; chromosome 18 deletions in 73% of carcinomas, 47% of advanced adenomas; 17p loss in 75% of carcinomas; alterations accumulated in parallel with clinical progression. The 1990 Cell paper (Fearon, Vogelstein), titled as a genetic model for colorectal tumorigenesis, has no PubMed abstract and was not read, so its content is not summarised.
  2. 1990 · Treatment · Chemotherapy⚑ needs reviewAdjuvant fluorouracil plus levamisole for stage III colon cancerIn a large randomised trial of people with resected colon cancer, one year of fluorouracil plus levamisole lowered the risk of recurrence by about forty percent in cancer that had reached the lymph nodes, compared with observation, and reduced deaths by about a third. The authors concluded this should be standard treatment for that stage.Moertel 1990: 1296 patients with resected stage B2 or C colon cancer randomised to observation or 1 year of levamisole plus fluorouracil (stage C also to levamisole alone). Median follow-up 3 years. In stage C, levamisole plus fluorouracil reduced recurrence risk by 41% (P<0.0001) and the death rate by 33% (P approx. 0.006); levamisole alone had no detectable effect; stage B2 results were equivocal. Final report 1995: 929 eligible stage III patients, 40% and 33% reductions.
  3. 1993 · Detection · Detection & preventionFaecal occult blood screening trials and mortalityIn a Minnesota trial, people offered a yearly stool test for hidden blood had fewer deaths from bowel cancer over 13 years than people not offered it; testing every two years did not differ significantly. Trials in England and Denmark also reported fewer deaths. The tests, intervals and uptake differed, so each trial is read against its own control group.Randomised faecal occult blood test (FOBT) trials versus unscreened controls. Minnesota (46,551 aged 50-80): 13-year cumulative colorectal cancer mortality per 1,000 was 5.88 annual (about 83% of slides rehydrated), 8.33 biennial, 8.83 control; only annual was significantly lower. At 30 years, relative risk 0.68 annual, 0.78 biennial; all-cause mortality unchanged (1.00). Nottingham (biennial, not rehydrated): 360 vs 420 deaths, odds ratio 0.85 (95% CI 0.74-0.98). Funen: mortality ratio 0.82 (0.68-0.99).
  4. 1993 · Understanding · UnderstandingLynch syndrome and DNA mismatch repair genesFamilies with many bowel, stomach and womb cancers had been recorded since 1895. In 1993 and 1994, studies linked such families to a region of chromosome 2, to errors in short repeated stretches of tumour DNA, and to genes that normally repair DNA copying mistakes. The condition is now called Lynch syndrome.Papers from 1993 and 1994: familial colorectal cancer linked to chromosome 2 (lod scores 6.39 and 1.45 in two kindreds); widespread microsatellite instability in familial tumours (13% of sporadic cancers shared it); instability in 25 of 90 tumours (28%), more often proximal colon; hMSH2 and hMLH1 proposed as HNPCC genes. Hampel 2008: 44 of 1,566 patients with colorectal cancer (2.8%, 95% CI 2.1-3.8) had Lynch syndrome.
  5. 1993 · Detection · Detection & preventionPolypectomy and colorectal cancer prevention (National Polyp Study)The National Polyp Study followed people who had adenomas removed at colonoscopy. Fewer colorectal cancers appeared than expected from comparison groups, and a later report found fewer deaths from the disease than expected in the general population. The comparison was with expected rates, not a randomised control group.National Polyp Study (cohort, not randomised). 1993: 1,418 people with adenomas removed, average follow-up 5.9 years (8,401 person-years); 5 asymptomatic early-stage cancers vs 48.3, 43.4 and 20.7 expected from three reference groups (reductions 90, 88 and 76 percent, P<0.001). 2012: 2,602 people, median follow-up 15.8 years; 12 colorectal cancer deaths vs 25.4 expected, standardized incidence-based mortality ratio 0.47 (95% CI 0.26-0.80).
  6. 1995 · Treatment · Chemotherapy⚑ needs reviewAdjuvant fluorouracil with folinic acid: IMPACT and QUASAR (stage II)In a 1995 pooled analysis of three randomised trials, three-year survival after surgery for Dukes' B and C colon cancer was about 83 in 100 with six months of fluorouracil and folinic acid, against about 78 in 100 in the control group. QUASAR (2007) tested the same drugs in mostly stage II disease; its authors called the absolute gain small.IMPACT (3 randomised trials pooled; 1493 eligible, Dukes' B and C): fluorouracil plus folinic acid reduced mortality by 22% (95% CI 3-38; p=0.029) and events by 35%; 3-year EFS 62% to 71%, OS 78% to 83%. IMPACT B2 (1016 patients, five trials): 5-year EFS 73% vs 76%, OS 80% vs 82%, not significant. QUASAR (3239; 91% stage II): relative risk of death 0.82 (95% CI 0.70-0.95; p=0.008).

Targeted therapy

  1. 2000 · Treatment · ChemotherapyIrinotecan with fluorouracil as first-line therapy for metastatic CRCTwo randomised trials in 2000 added irinotecan to fluorouracil and leucovorin as the first treatment for bowel cancer that had spread. In both, median survival was longer than with fluorouracil and leucovorin alone: 17.4 versus 14.1 months in one trial and 14.8 versus 12.6 months in the other.Douillard (Lancet, 25 March 2000; n=387): irinotecan plus infusional fluorouracil/folinate vs fluorouracil/folinate alone; median OS 17.4 vs 14.1 months (p=0.031); TTP 6.7 vs 4.4 months. FDA approved first-line use 20 April 2000. Saltz (NEJM, 28 September 2000; n=683, ITT): IFL vs bolus fluorouracil/leucovorin; median OS 14.8 vs 12.6 months (P=0.04), PFS 7.0 vs 4.3 months (P=0.004), response 39% vs 21%.
  2. 2000 · Treatment · ChemotherapyOxaliplatin added to infusional fluorouracil in advanced CRCTwo randomised trials in 2000 added oxaliplatin to fluorouracil and leucovorin as the first treatment for advanced bowel cancer. Tumours responded more than twice as often with oxaliplatin and took longer to progress, but overall survival was not significantly better in the larger trial and was similar in the other, and nerve damage and diarrhoea were more frequent.Giacchetti (n=200; chronomodulated fluorouracil-leucovorin): response 53% vs 16% (P<.001); median PFS 8.7 vs 6.1 months (P=.048); median OS 19.4 vs 19.9 months. de Gramont (n=420; LV5FU2): median PFS 9.0 vs 6.2 months (P=.0003); response 50.7% vs 22.3%; median OS 16.2 vs 14.7 months (P=.12). FDA accelerated approval 9 August 2002; first-line indication 9 January 2004.
  3. 2003 · Understanding · Understanding⚑ needs reviewMismatch repair status and benefit from adjuvant fluorouracilRetrospective analyses of tumour samples from earlier randomised trials suggested that colon cancers with microsatellite instability or mismatch repair deficiency did not benefit from adjuvant fluorouracil, while other colon cancers did. They were not new randomised comparisons, the second report also pooled the first one's patients, and the regimens contained no oxaliplatin.Retrospective biomarker analyses of specimens from earlier randomised fluorouracil trials (stage II and III). Ribic 2003 (570 specimens; 16.7% high-frequency MSI): chemotherapy OS HR 0.72 (95% CI 0.53-0.99) in stable or low-frequency MSI tumours, no benefit in high-frequency MSI; interaction P=0.01. Sargent 2010 (457 new patients; pooled with Ribic's 570: 1027, 165 dMMR): DFS HR 0.67 (0.48-0.93) in pMMR; in dMMR 1.10 (0.42-2.91) in the abstract, 1.39 (0.46-4.15) in the full text.
  4. 2004 · Treatment · SurgeryLaparoscopic versus open colectomy: COST, CLASICC, COLOR, JCOG0404Four randomised trials compared laparoscopic-assisted (keyhole) surgery with open surgery: three in colon cancer and CLASICC in colon and rectal cancer together. In COST, about 16 in 100 had a recurrence by three years with laparoscopic surgery and 18 in 100 with open surgery. COLOR could not formally show non-inferiority, and JCOG0404 did not meet its survival margin.COST (872, colon): 3-year recurrence 16% (laparoscopic) vs 18% (open); HR 0.86 (95% CI 0.63-1.17). COLOR (1248 randomised): 3-year DFS 74.2% (laparoscopic) vs 76.2% (open), difference 2.0% (95% CI -3.2 to 7.2) against a 7% non-inferiority boundary. JCOG0404 (1057, stage II or III): 5-year OS 91.8% (laparoscopic) vs 90.4% (open), HR 1.06 (90% CI 0.79-1.41), non-inferiority not met. CLASICC (794, colon and rectum): Dukes' C2 6% (laparoscopic) vs 7% (open).
  5. 2004 · Treatment · ChemotherapyMOSAIC: oxaliplatin added to adjuvant fluorouracil for colon cancerIn MOSAIC, people who had surgery for stage II or III colon cancer were randomly given six months of fluorouracil with leucovorin, with or without oxaliplatin. At three years about seventy-eight in a hundred were alive without recurrence with oxaliplatin, against about seventy-three without. At ten years, survival was higher with oxaliplatin in stage III but not stage II.MOSAIC (André 2004): 2246 patients with resected stage II or III colon cancer randomised to LV5FU2 or FOLFOX4 for 6 months. Median follow-up 37.9 months: 3-year DFS 78.2% (95% CI 75.6-80.7) vs 72.9% (70.2-75.7); HR for recurrence 0.77, P=0.002. At 10 years, stage III OS 67.1% vs 59.0% (HR 0.80, P=0.016); stage II OS HR 1.00 (P=0.980).
  6. 2004 · Treatment · TargetedCetuximab and panitumumab: anti-EGFR antibodies in refractory CRCCetuximab and panitumumab are antibodies that block EGFR, a growth-signal receptor on tumour cells. In bowel cancer that had stopped responding to chemotherapy, cetuximab was active alone and with irinotecan, and both drugs delayed progression compared with supportive care alone. The FDA approved them in 2004 and 2006 under its accelerated pathway.BOND (n=329, irinotecan-refractory): response 22.9% with cetuximab plus irinotecan vs 10.8% with cetuximab alone; median TTP 4.1 vs 1.5 months; median survival 8.6 vs 6.9 months (P=0.48). CO.17 (n=572): cetuximab plus BSC vs BSC, OS HR 0.77 (95% CI 0.64 to 0.92), median 6.1 vs 4.6 months. Panitumumab plus BSC vs BSC (n=463): PFS HR 0.54 (0.44 to 0.66); OS HR 1.00 (0.82 to 1.22). FDA accelerated approvals 12 February 2004 and 27 September 2006.
  7. 2004 · Treatment · TargetedBevacizumab added to first-line irinotecan-based chemotherapyIn 813 people with untreated bowel cancer that had spread, adding bevacizumab, a drug that targets the growth of blood vessels, to a chemotherapy regimen containing irinotecan lengthened median survival from 15.6 to 20.3 months. The FDA approved it for first-line use in February 2004.813 patients with previously untreated metastatic CRC randomised to IFL plus bevacizumab 5 mg/kg every 2 weeks (n=402) or IFL plus placebo (n=411). Primary endpoint OS: median 20.3 vs 15.6 months, HR 0.66, P<0.001. PFS 10.6 vs 6.2 months (HR 0.54, P<0.001); response rate 44.8% vs 34.8% (P=0.004). Grade 3 hypertension 11.0% vs 2.3%. FDA approval 26 February 2004.
  8. 2005 · Treatment · ChemotherapyX-ACT: adjuvant capecitabine for stage III colon cancerIn X-ACT, people with stage III colon cancer were randomly given 24 weeks of capecitabine tablets or bolus fluorouracil with leucovorin by injection after surgery. Capecitabine was at least as good at keeping the cancer from returning and had fewer side effects. The EU (March 2005) and the US (June 2005) extended capecitabine's approval to this use.X-ACT: 1987 patients with resected stage III colon cancer randomised to 24 weeks of capecitabine (1004) or bolus fluorouracil plus leucovorin, Mayo regimen (983). Primary aim at least equivalence in DFS (margin 1.20; met). 2005: RFS HR 0.86 (95% CI 0.74-0.99; P=0.04), fewer adverse events (P<0.001). Final (median 6.9 years): DFS HR 0.88 (0.77-1.01), OS HR 0.86 (0.74-1.01). FDA, June 2005: non-inferior to 5-FU/LV for DFS.
  9. 2006 · De-escalation · ChemotherapyStop-and-go and maintenance therapy: OPTIMOX1 to PRODIGE 9Trials asked whether people with spread bowel cancer could take planned breaks from chemotherapy or use lighter maintenance treatment. Oxaliplatin could be stopped after six cycles without a clear survival loss in OPTIMOX1. Planned breaks did not meet the survival target in COIN. Capecitabine with bevacizumab maintenance delayed second progression; bevacizumab alone after FOLFIRI did not improve outcomes.OPTIMOX1 (n=620): FOLFOX7 stop-and-go vs FOLFOX4, median PFS 8.7 vs 9.0 months, OS 21.2 vs 19.3 (P not significant). COIN intermittent vs continuous: median OS 14.4 vs 15.8 months (HR 1.084); non-inferiority not shown. CAIRO3 (n=558): maintenance capecitabine-bevacizumab vs observation, PFS2 11.7 vs 8.5 months (HR 0.67, 95% CI 0.56-0.81). AIO 0207: bevacizumab alone non-inferior on time to failure of strategy (HR 1.08). PRODIGE 9 (bevacizumab maintenance vs observation): median OS 21.7 vs 22.0 months.
  10. 2007 · Treatment · ChemotherapyFOLFOXIRI with bevacizumab: GONO trial, TRIBE and TRIBE2Using three chemotherapy drugs together instead of two, with bevacizumab, delayed progression of untreated bowel cancer that had spread by about two and a half months in TRIBE; an update reported about four months longer median survival. Side effects were more common, and the trials enrolled people up to age 75 who met performance-status criteria.GONO (n=244): FOLFOXIRI vs FOLFIRI, response 66% vs 41% (primary endpoint), median OS 22.6 vs 16.7 months (HR 0.70). TRIBE (n=508): with bevacizumab, median PFS 12.1 vs 9.7 months (HR 0.75, 95% CI 0.62-0.90; primary endpoint); OS 31.0 vs 25.8 months (HR 0.79, P=0.054), updated 29.8 vs 25.8 months (HR 0.80, 0.65-0.98). TRIBE2 (n=679): FOLFOXIRI-bevacizumab with reintroduction vs mFOLFOX6-bevacizumab then FOLFIRI-bevacizumab, PFS2 19.2 vs 16.4 months (HR 0.74).
  11. 2008 · Reversal · TargetedKRAS mutation predicts no benefit: anti-EGFR limited to wild-typeCetuximab and panitumumab were first approved for bowel cancers carrying the EGFR receptor. In 2008, two analyses of randomised trials showed that people whose tumours had a KRAS mutation gained no benefit, while those without it did. The FDA later revised the cetuximab label and then limited its use to tumours without the mutation.Biomarker analyses of two randomised trials of anti-EGFR antibody plus best supportive care versus best supportive care alone. Panitumumab (KRAS known in 427 of 463): PFS HR 0.45 (95% CI 0.34 to 0.59) in wild-type vs 0.99 (0.73 to 1.36) in mutant tumours. Cetuximab, CO.17 (K-ras tested in 394 of 572): wild-type OS median 9.5 vs 4.8 months, HR 0.55 (0.41 to 0.74); mutant OS HR 0.98 (P=0.89). FDA limited cetuximab to K-ras wild-type disease on 6 July 2012.
  12. 2008 · Treatment · ChemotherapyChemotherapy around liver resection: EORTC 40983, New EPOC, JCOG0603For people whose bowel cancer had spread only to the liver and could be removed, trials tested chemotherapy around surgery. FOLFOX before and after surgery delayed progression in eligible and resected patients, but overall survival did not differ later. Adding cetuximab was linked to shorter survival. After surgery, chemotherapy delayed recurrence in JCOG0603, but long-term survival did not differ.EORTC 40983 (364 patients, resectable liver metastases): perioperative FOLFOX4 vs surgery alone, 3-year PFS 35.4% vs 28.1% (HR 0.79, P=0.058, ITT); long-term OS HR 0.88 (95% CI 0.68-1.14). New EPOC (KRAS wild-type): chemotherapy with vs without cetuximab, long-term median OS 55.4 vs 81.0 months (HR 1.45). JCOG0603 (hepatectomy plus mFOLFOX6 vs hepatectomy alone): 5-year DFS 49.7% vs 40.5% (HR 0.72); 5-year OS 73.4% vs 80.1% (HR 1.07, 0.73-1.57).
  13. 2010 · Reversal · TargetedAdjuvant bevacizumab and cetuximab trials in colon cancerFour large randomised trials added a targeted drug, bevacizumab or cetuximab, to oxaliplatin-based chemotherapy after surgery for colon cancer. None improved disease-free survival. In one trial the authors noted overall survival data suggesting a possible harm with bevacizumab, and cetuximab added side effects. Both drugs had shown benefit in advanced disease, which is a different setting.Adjuvant colon cancer, DFS primary endpoint. NSABP C-08 (bevacizumab + mFOLFOX6; 2672 analysed): HR 0.89 (95% CI 0.76-1.04), P=0.15. AVANT (3451 randomised; stage III primary analysis): HR 1.17 (0.98-1.39) bevacizumab-FOLFOX4 and 1.07 (0.90-1.28) bevacizumab-XELOX vs FOLFOX4; OS HR 1.27 (1.03-1.57) for bevacizumab-FOLFOX4. N0147 (KRAS wild-type, mFOLFOX6 with or without cetuximab): 3-year DFS 71.5% with vs 74.6% without, HR 1.21 (0.98-1.49). PETACC-8 (KRAS exon 2 wild-type, 1602): HR 1.05 (0.85-1.29).
  14. 2010 · Treatment · TargetedFirst-line anti-EGFR trials and extended RAS testing (PRIME, FIRE-3)In people with untreated bowel cancer that had spread, adding an antibody against the EGF receptor to chemotherapy was linked to better outcomes when the tumour had no RAS gene changes, not when it had them. In one trial's RAS analysis, median survival with panitumumab was about six months longer. Testing widened from KRAS exon 2 to KRAS and NRAS.PRIME (n=1,183, FOLFOX4 with or without panitumumab): KRAS wild-type median PFS 9.6 vs 8.0 months (HR 0.80, 95% CI 0.66-0.97); prospective-retrospective RAS wild-type analysis (512 patients) median OS 26.0 vs 20.2 months (HR 0.78, 0.62-0.99). Other RAS mutations (17% of KRAS exon 2 wild-type tumours) were associated with inferior outcomes with panitumumab. FIRE-3 (FOLFIRI, KRAS exon 2 wild-type), cetuximab vs bevacizumab: response 62.0% vs 58.0% (primary endpoint, P=0.18); median OS 28.7 vs 25.0 months (HR 0.77, not primary).
  15. 2010 · Detection · Detection & preventionFlexible sigmoidoscopy screening trialsFour randomised trials, in the UK, Italy, the US and Norway, offered flexible sigmoidoscopy (an examination of the lower bowel) or no screening. All four reported fewer bowel cancers in the invited group. Bowel cancer deaths were significantly lower in three; the Italian difference was not significant. Later Norwegian data showed reductions in men, little or no effect in women.Colorectal cancer mortality by randomised group in four sigmoidoscopy trials: UK once-only, HR 0.69 (95% CI 0.59-0.82), median 11.2 years; NORCCAP, HR 0.73 (0.56-0.94); PLCO (repeat screening), RR 0.74 (0.63-0.87); SCORE, RR 0.78 (0.56-1.08, not significant). PLCO distal cancer death RR 0.50, proximal 0.97. NORCCAP at 14.8 years: men HR 0.63 (0.47-0.83), women 1.01 (0.77-1.33).
  16. 2010 · Detection · Detection & preventionAdenoma detection rate and interval cancer after colonoscopyStudies of screening colonoscopy found that colorectal cancers diagnosed after a colonoscopy (interval cancers) were less common among patients of endoscopists who found adenomas more often. In one large health-system study, patients of the highest-detecting physicians had about half the interval cancer risk of those of the lowest-detecting. These are observational findings, not trial results.Observational cohorts linking endoscopist adenoma detection rate (ADR) to interval cancer, not trials. Kaminski 2010: 186 endoscopists, 45,026 subjects, 42 interval cancers; ADR below 11.0% vs 20.0% or higher, HR 10.94 (95% CI 1.37-87.01); cecal intubation rate not associated. Corley 2014: 314,872 colonoscopies, 136 gastroenterologists; highest vs lowest ADR quintile, interval cancer HR 0.52 (0.39-0.69), fatal 0.38 (0.22-0.65); per 1.0% ADR increase, HR 0.97.

Immunotherapy

  1. 2012 · Treatment · TargetedRegorafenib for previously treated metastatic CRC (CORRECT)For people whose bowel cancer had kept growing after all standard treatments, a placebo-controlled trial tested regorafenib, a tablet that blocks several growth signals. Survival was longer than with placebo, by a little over a month, at the cost of frequent side effects such as hand-foot skin reaction.Double-blind phase 3 trial, 760 patients with metastatic colorectal cancer progressing during or within 3 months after the last standard therapy, randomised 2:1 to regorafenib (n=505) or placebo (n=255), both with best supportive care. Median OS 6.4 vs 5.0 months (HR 0.77, 95% CI 0.64-0.94; one-sided P=0.0052; preplanned interim analysis). Hand-foot skin reaction of grade 3 or higher: 17% with regorafenib.
  2. 2014 · Treatment · TargetedASPECCT: panitumumab vs cetuximab in refractory KRAS wild-type mCRCASPECCT compared two EGFR-blocking antibodies, panitumumab and cetuximab, in people whose bowel cancer had progressed on chemotherapy and had no KRAS exon 2 mutation. Median survival was about ten months with each, and panitumumab met the trial's non-inferiority criterion. The FDA's 2014 supplement letter lists this study, by sponsor study number, among those used to verify panitumumab's clinical benefit.ASPECCT: 1010 randomised 1:1, 999 treated (panitumumab 499, cetuximab 500); chemotherapy-refractory KRAS exon 2 wild-type mCRC. Primary OS non-inferiority met: median 10.4 vs 10.0 months, HR 0.97 (95% CI 0.84 to 1.11), Z=-3.19, P=0.0007; panitumumab retained 105.7% of the cetuximab effect. Final analysis: OS 10.2 vs 9.9 months, HR 0.94 (0.82 to 1.07); PFS 4.2 vs 4.4 months, HR 0.98 (0.87 to 1.12).
  3. 2014 · Detection · Detection & preventionMultitarget stool DNA and RNA screening tests (FDA 2014 to 2024)Stool tests that look for altered DNA or RNA plus blood were compared with colonoscopy, and with an established stool blood test (FIT), in thousands of adults. They found more cancers and precancerous lesions than the FIT but also gave more false alarms. The US FDA approved all three tests studied: Cologuard (2014), ColoSense and Cologuard Plus (2024).Paired diagnostic-accuracy studies against colonoscopy, each with a FIT comparator; none has an incidence or mortality outcome. DeeP-C (9,989 evaluable): sensitivity for colorectal cancer 92.3% vs 73.8% with FIT (P=0.002); specificity for advanced neoplasia 86.6% vs 94.9%. BLUE-C (20,176): 93.9% (95% CI 87.1-97.7) vs 67.3%; specificity 90.6% vs 94.8%. CRC-PREVENT (8,920): 94.4% vs 77.8%. FDA approvals: Cologuard 2014, ColoSense and Cologuard Plus 2024.
  4. 2015 · Treatment · ImmunotherapyPD-1 blockade in mismatch repair-deficient colorectal cancerTumours that cannot repair DNA copying errors carry many mutations. In a small 2015 study, pembrolizumab produced responses in 4 of 10 such bowel cancers and in none of 18 without the repair defect. Later nivolumab studies, with and without ipilimumab, reported responses in about a third to over half of previously treated patients, all without a comparison group.Non-randomised phase 2 studies. Le 2015 (n=41 across cohorts): immune-related ORR 40% (4 of 10) in dMMR colorectal vs 0% (0 of 18) in pMMR. CheckMate 142, previously treated: nivolumab ORR 31.1% (95% CI 20.8 to 42.9; n=74); nivolumab plus ipilimumab 55% (45.2 to 63.8; n=119). First line, nivolumab plus low-dose ipilimumab: 69% (53 to 82; n=45). FDA accelerated approvals 31 July 2017 and 10 July 2018.
  5. 2015 · Treatment · ChemotherapyTrifluridine-tipiracil (TAS-102): RECOURSE and SUNLIGHTTrifluridine-tipiracil is an oral chemotherapy tablet. In a placebo-controlled trial of refractory metastatic bowel cancer it lengthened survival by under two months. A later trial in a similar setting found that adding bevacizumab to it lengthened survival by about three months compared with the tablet alone.RECOURSE (double-blind phase 3, n=800, 2:1): median OS 7.1 months with trifluridine-tipiracil vs 5.3 with placebo (HR 0.68, 95% CI 0.58-0.81; P<0.001). SUNLIGHT (open-label phase 3, 246 per arm, at most two prior regimens): trifluridine-tipiracil plus bevacizumab vs trifluridine-tipiracil alone, median OS 10.8 vs 7.5 months (HR 0.61, 95% CI 0.49-0.77); median PFS 5.6 vs 2.4 months (HR 0.44, 95% CI 0.36-0.54).
  6. 2016 · Treatment · TargetedHER2-targeted therapy for HER2-positive metastatic CRCA small share of bowel cancers carry extra HER2, the target of breast-cancer drugs. Four separate studies without a comparison group gave HER2-directed drugs to people whose cancer had stopped responding to standard treatment; each reported tumour responses in roughly 30 to 45 of every 100 people treated. The US provisionally approved tucatinib with trastuzumab in 2023.Four single-arm phase 2 studies in HER2-positive or HER2-amplified metastatic colorectal cancer, each reporting its own objective response rate (not comparable across studies): HERACLES trastuzumab + lapatinib 30% (8 of 27; 95% CI 14-50); MyPathway pertuzumab + trastuzumab 32% (18 of 57; 95% CI 20-45); DESTINY-CRC01 trastuzumab deruxtecan 45.3% (cohort A, n=53; 95% CI 31.6-59.6); MOUNTAINEER tucatinib + trastuzumab 38.1% (n=84; 95% CI 27.7-49.3).
  7. 2017 · Understanding · TargetedPrimary tumour side and first-line biologic: 80405, FIRE-3, PARADIGMWhere in the bowel a cancer started, left or right, was linked to outcomes and to how well anti-EGFR antibodies worked. A pooled look back at six trials found the clearest antibody benefit in left-sided tumours. In PARADIGM, which tested survival in left-sided tumours first, median survival was about three and a half months longer with panitumumab than with bevacizumab.Retrospective pooled analysis of six randomised trials (RAS wild-type; side and RAS status available for 2,159 of 5,760 patients): EGFR antibody vs control, OS HR 0.75 (95% CI 0.67-0.84) in left-sided and 1.12 (0.87-1.45) in right-sided tumours (interaction P<0.001). CALGB/SWOG 80405 (KRAS wild-type, n=1,137): median OS 30.0 vs 29.0 months, cetuximab vs bevacizumab (HR 0.88, 0.77-1.01). PARADIGM (RAS wild-type; OS tested first in left-sided tumours): 37.9 vs 34.3 months, panitumumab vs bevacizumab (HR 0.82).
  8. 2018 · Treatment · TargetedFruquintinib for refractory metastatic CRC (FRESCO and FRESCO-2)Fruquintinib is a capsule that blocks the receptors for a blood-vessel growth signal. A trial in China and a larger international trial each compared it with placebo in people whose bowel cancer had progressed after standard treatments, and each found longer survival with fruquintinib.FRESCO (double-blind phase 3, China, n=416, 2:1): median OS 9.3 vs 6.6 months, fruquintinib vs placebo (HR 0.65, 95% CI 0.51-0.83; P<0.001). FRESCO-2 (double-blind phase 3, 14 countries, n=691, 2:1, progressed on or intolerant to trifluridine-tipiracil or regorafenib or both): median OS 7.4 vs 4.8 months (HR 0.66, 95% CI 0.55-0.80; P<0.0001). Grade 3 or worse adverse events 63% vs 50% in FRESCO-2.
  9. 2018 · De-escalation · ChemotherapyIDEA: 3 versus 6 months of adjuvant oxaliplatin chemotherapySix randomised trials were pooled to ask whether three months of oxaliplatin chemotherapy after surgery for stage III colon cancer works as well as six months, with less nerve damage. Overall, the pooled analysis could not confirm that three months was as good. For the CAPOX regimen it met the standard set in advance, but for FOLFOX it did not.IDEA pooled analysis (Grothey 2018): six phase 3 trials, 12,834 patients with stage III colon cancer, 3 vs 6 months of FOLFOX or CAPOX. Noninferiority (upper CI limit 1.12) was not confirmed overall (HR 1.07, 95% CI 1.00-1.15); seen for CAPOX (HR 0.95, 0.85-1.06), not FOLFOX (HR 1.16, 1.06-1.26). Final 5-year OS 82.4% vs 82.8% (HR 1.02, 0.95-1.11; noninferiority p=0.058).
  10. 2018 · Understanding · UnderstandingACS lowers screening start age to 45 amid rising early-onset incidenceIn 2018 the American Cancer Society lowered its advised starting age for colorectal screening in average-risk adults to 45; the US Preventive Services Task Force followed in 2021. Registry data show colorectal cancer becoming more common in adults under 50, in the US and many other countries, while it has fallen in older adults. The cause is still being studied.Registry analyses and guideline statements, not trials. SEER areas 1974-2013 (490,305 cases): colon cancer incidence rose 1.0% to 2.4% annually in ages 20-39 and 0.5% to 1.3% in ages 40-54. A 50-country registry analysis found early-onset incidence rising in 27 of 50 countries. In the 2026 ACS statistics overall incidence fell 0.9% annually in 2013-2022 while it rose 3% annually in ages 20-49. ACS 2018: start at 45 (qualified recommendation); USPSTF 2021: grade B at ages 45-49.
  11. 2019 · Treatment · TargetedBEACON: encorafenib plus cetuximab for BRAF V600E mCRCSome bowel cancers carry a change called BRAF V600E and do poorly on standard chemotherapy. In a randomised trial of people whose cancer had progressed after one or two earlier treatments, adding encorafenib to cetuximab lengthened survival by several months. The US regulator approved the two-drug version, not the three-drug one tested alongside it.Open-label phase 3 trial, 665 patients with BRAF V600E-mutated metastatic colorectal cancer after one or two prior regimens, randomised 1:1:1 to triplet (encorafenib, binimetinib, cetuximab), doublet (encorafenib, cetuximab) or control (cetuximab with irinotecan or FOLFIRI). Interim median OS 9.0 vs 5.4 months, triplet vs control (HR 0.52, 95% CI 0.39-0.70); doublet 8.4 months (HR 0.60, 95% CI 0.45-0.79). Confirmed ORR 26% vs 2%.

Precision & engineered

  1. 2020 · Treatment · ImmunotherapyKEYNOTE-177: pembrolizumab first-line for MSI-H/dMMR metastatic CRCIn untreated advanced MSI-H/dMMR bowel cancer (tumours that cannot repair DNA copying errors), pembrolizumab was compared with chemotherapy. Cancer progressed after a median of 16.5 versus 8.2 months, with fewer severe side effects. At over five years median survival was 77.5 versus 36.7 months, but the final survival analysis was not statistically significant and many chemotherapy patients later received immunotherapy.Open-label phase 3: 307 patients with untreated MSI-H/dMMR metastatic CRC randomised 1:1 to pembrolizumab (n=153) or investigator-choice chemotherapy (n=154), crossover allowed. PFS median 16.5 vs 8.2 months (HR 0.60, 95% CI 0.45 to 0.80; P=0.0002; second interim analysis). Final OS HR 0.74 (95% CI 0.53 to 1.03; p=0.036), prespecified significance not met. At over 5 years, OS HR 0.73 (95% CI 0.53 to 0.99); effective crossover 62%. FDA approval 29 June 2020.
  2. 2021 · Reversal · SurgeryPRODIGE 7: HIPEC added to cytoreductive surgery (peritoneal CRC)PRODIGE 7 randomly assigned 265 people with colorectal cancer that had spread to the lining of the abdomen to surgery that removes all visible tumour, with or without heated chemotherapy washed through the abdomen. Median survival was about forty-one to forty-two months in both groups, and later complications were more frequent with the heated chemotherapy.PRODIGE 7 (Quénet 2021): phase 3, 265 patients with colorectal peritoneal metastases (PCI 25 or less) after complete cytoreduction, randomised to surgery plus oxaliplatin HIPEC (n=133) or surgery alone (n=132). Median OS 41.7 months (95% CI 36.2-53.8) vs 41.2 (35.1-49.7); HR 1.00 (95.37% CI 0.63-1.58), P=0.99; median follow-up 63.8 months. Grade 3 or worse adverse events at 60 days: 26% vs 15% (P=0.035).
  3. 2022 · Treatment · ChemotherapyDYNAMIC: ctDNA-guided adjuvant therapy in stage II colon cancerIn DYNAMIC, people with stage II colon cancer were randomly assigned to have chemotherapy decisions guided by a blood test for circulating tumour DNA or by standard features of the tumour. Fewer people in the test-guided group had chemotherapy, about fifteen in a hundred against twenty-eight, and two-year recurrence-free survival was about the same, near ninety-three percent.DYNAMIC (Tie 2022): 455 patients with stage II colon cancer randomised 2:1 to ctDNA-guided (n=302) or standard management (n=153). Adjuvant chemotherapy 15% vs 28%. 2-year RFS 93.5% vs 92.4% (difference 1.1 points, 95% CI -4.1 to 6.2; noninferiority margin -8.5), median follow-up 37 months. 5-year RFS 88% vs 87%; 5-year OS 93.8% vs 93.3% (HR 1.05, P=0.887). DYNAMIC-III (stage III): de-escalation did not meet its noninferiority margin.
  4. 2022 · Treatment · Targeted⚑ needs reviewKRAS G12C inhibitors with anti-EGFR antibodiesA small share of bowel cancers carry a KRAS G12C change. Used alone, drugs aimed at it worked modestly. Paired with an anti-EGFR antibody, one drug shrank tumours in nearly half of 28 evaluable people in a non-randomised study; its US approval was withdrawn in 2026. Another delayed progression against standard treatment in a randomised trial.KRYSTAL-1 (phase 1-2, nonrandomised, KRAS G12C colorectal cancer): adagrasib + cetuximab, ORR 46% (95% CI 28-66; 28 evaluable patients), median PFS 6.9 months (32 patients). CodeBreaK 300 (phase 3, n=160): sotorasib 960 mg + panitumumab vs investigator's choice, median PFS 5.6 vs 2.2 months (HR 0.49, 95% CI 0.30-0.80; P=0.006); ORR 26.4% vs 0%. OS HR 0.70 (95% CI 0.41-1.18; P=0.20).
  5. 2022 · Detection · Detection & preventionNordICC: invitation to screening colonoscopyIn a trial in four European countries (follow-up data from three), people aged 55 to 64 were randomly invited to one screening colonoscopy or not invited. Fewer than half of those invited had the examination. After about ten years, fewer colorectal cancers were found in the invited group; deaths from colorectal cancer were not clearly different, with an imprecise estimate.Pragmatic randomised trial of invitation to one screening colonoscopy vs usual care (ages 55-64; 84,585 with follow-up in Poland, Norway, Sweden); 42.0% of the invited group were screened. Intention-to-screen at 10 years: colorectal cancer risk 0.98% vs 1.20% (risk ratio 0.82, 95% CI 0.70-0.93); colorectal cancer death 0.28% vs 0.31% (risk ratio 0.90, 0.64-1.16); all-cause death risk ratio 0.99 (0.96-1.04). Separate instrumental-variable per-protocol analysis: incidence reduction 35 to 41%; mortality intervals include 1.
  6. 2023 · Treatment · ChemotherapyFOxTROT: neoadjuvant chemotherapy for locally advanced colon cancerIn FOxTROT, people with colon cancer growing into or beyond the bowel wall were randomly given six weeks of chemotherapy before surgery and eighteen weeks after, or twenty-four weeks after surgery alone. Fewer in the before-surgery group had residual or returned disease within two years, about seventeen in a hundred against twenty-two. Benefit looked small with mismatch repair deficiency.FOxTROT (Morton 2023): 1053 patients (699 NAC, 354 control) with radiologically staged T3-4 N0-2 M0 colon cancer randomised 2:1 to 6 weeks of preoperative oxaliplatin-fluoropyrimidine plus 18 weeks postoperatively, or 24 weeks postoperatively. Residual disease or recurrence within 2 years: 16.9% vs 21.5% (rate ratio 0.72, 95% CI 0.54-0.98, P=0.037); complete resection 94% vs 89%. Panitumumab added no benefit.
  7. 2024 · Treatment · ImmunotherapyCheckMate 8HW: nivolumab plus ipilimumab in MSI-H/dMMR mCRCColorectal cancers with faulty DNA mismatch repair can respond to immune checkpoint drugs. In a randomised trial, people given nivolumab plus ipilimumab as their first treatment stayed free of progression longer than those given chemotherapy. A second report found they also stayed free of progression longer than with nivolumab alone. In 2025 the US approvals became full approvals.Open-label phase 3 trial, 2:2:1 randomisation (nivolumab + ipilimumab, nivolumab, chemotherapy) in MSI-H/dMMR metastatic colorectal cancer. First line (NEJM 2024): 24-month PFS 72% vs 14% for nivolumab + ipilimumab vs chemotherapy (P<0.001; centrally confirmed, n=255). All lines (Lancet 2025): nivolumab + ipilimumab vs nivolumab, PFS HR 0.62 (95% CI 0.48-0.81; P=0.0003).
  8. 2024 · Treatment · ImmunotherapyNICHE-2: neoadjuvant nivolumab plus ipilimumab for dMMR colon cancerIn NICHE-2, 115 people with locally advanced colon cancer with mismatch repair deficiency had nivolumab and ipilimumab before surgery. Nearly all had surgery on time. Among 111 assessed, about two in three had no viable tumour left, and none had recurrence by a median of 26 months. It was a single-group study without a comparison arm.NICHE-2 (Chalabi 2024): phase 2, single-arm, 115 patients with untreated locally advanced dMMR colon cancer given neoadjuvant nivolumab plus ipilimumab. Timely surgery in 113 (98%; 97.5% CI 93-100). In 111 patients in the efficacy analysis, pathological response in 109 (98%; 95% CI 94-100), major pathological response in 105 (95%), pCR in 75 (68%). Grade 3-4 immune-related adverse events in 5 (4%). Median follow-up 26 months: no recurrence; 3-year DFS not reported in the abstract.
  9. 2024 · Treatment · Cell therapy & transplantTransMet: liver transplantation for unresectable liver metastasesIn a small, highly selected group of people whose bowel cancer had spread only to the liver and could not be removed, a European trial in 20 centres compared chemotherapy alone with chemotherapy plus a liver transplant. Five-year survival was about fifty-seven percent with a transplant and about thirteen percent without, but serious side effects were frequent.Open-label RCT in 94 patients with permanently unresectable colorectal liver metastases (resected BRAF-non-mutated primary, chemotherapy-responsive, no extrahepatic disease), 47 per arm: liver transplantation plus chemotherapy vs chemotherapy alone. ITT 5-year OS 56.6% (95% CI 43.2-74.1) vs 12.6% (5.2-30.1); HR 0.37 (0.21-0.65); P=0.0003. Per-protocol 73.3% vs 9.3%. Serious adverse events in 80% of transplanted patients; three retransplantations, one fatal.
  10. 2024 · Detection · Detection & preventionBlood-based screening tests: validation studies and FDA decisionsTwo blood tests that look for tumour DNA were each compared with colonoscopy in thousands of average-risk adults. In these studies they detected about four in five colorectal cancers but only about one in eight advanced precancerous lesions. The US FDA approved both tests, in 2024 and 2026, for adults aged 45 and older, with colonoscopy after a positive result.Prospective validation studies against screening colonoscopy in average-risk adults, not outcome trials. ECLIPSE (7,861 evaluable): sensitivity for colorectal cancer 83.1% (95% CI 72.2-90.3), advanced precancerous lesions 13.2%, specificity for advanced neoplasia 89.6%. PREEMPT CRC (27,010 evaluable): 79.2% (68.4-86.9), 12.5% (prespecified criterion not met), specificity 91.5%. FDA approved Shield (2024-07-26) and SimpleScreen CRC (2026-07-24) for adults aged 45 and older.
  11. 2024 · Treatment · TargetedBREAKWATER: first-line encorafenib, cetuximab and chemotherapyBowel cancers with a BRAF V600E change have had limited benefit from standard first-line chemotherapy. In a randomised trial, adding encorafenib and cetuximab to the chemotherapy mFOLFOX6 shrank more tumours, delayed progression and lengthened survival compared with standard care. US approval was provisional (accelerated) in 2024; in 2026 the regulator recorded its follow-up requirement as fulfilled.Open-label phase 3 trial, previously untreated BRAF V600E-mutant metastatic colorectal cancer: encorafenib + cetuximab + mFOLFOX6 vs standard care (chemotherapy with or without bevacizumab). ORR by blinded central review 60.9% vs 40.0% (odds ratio 2.443, one-sided P=0.0008). PFS median 12.8 vs 7.1 months (HR 0.53, 95% CI 0.41-0.68). Interim OS median 30.3 vs 15.1 months (HR 0.49, 95% CI 0.38-0.63).
  12. 2025 · Treatment · Detection & prevention⚑ needs reviewAdjuvant aspirin: ASCOLT and ALASCCA (PI3K-altered tumours)Two randomised trials tested daily aspirin after surgery for bowel cancer. ASCOLT did not require a particular tumour gene change and found no significant difference in disease-free survival. ALASCCA enrolled only tumours with PI3K-pathway gene changes; in its main group, recurrence by three years was about 8 in 100 with aspirin and 14 in 100 with placebo.ASCOLT (aspirin 200 mg vs placebo for 3 years after standard adjuvant therapy; 1587 randomised; Dukes' C or high-risk B colon, B/C rectal): 5-year DFS 77.0% vs 74.8%, HR 0.91 (95% CI 0.73-1.13; P=0.38). ALASCCA (aspirin 160 mg vs placebo for 3 years; PI3K-altered, stage I-III rectal or II-III colon; group A, PIK3CA exon 9/20 hotspot): 3-year recurrence 7.7% vs 14.1% (HR 0.49; 95% CI 0.24-0.98; P=0.04).
  13. 2025 · Treatment · Supportive careCHALLENGE: structured exercise after adjuvant chemotherapyIn CHALLENGE, people who had finished chemotherapy after colon cancer surgery were randomly given a three-year structured exercise programme or health-education materials alone. At five years about 80 in 100 with exercise were alive and free of recurrence or a new cancer, against about 74 in 100 with education alone. Muscle and joint problems were more common with exercise.CHALLENGE (phase 3, 55 centres): 889 patients with resected colon cancer who had completed adjuvant chemotherapy, randomised to a 3-year structured exercise programme (445) or health-education materials (444). Median follow-up 7.9 years: DFS (primary) HR 0.72 (95% CI 0.55-0.94; P=0.02); 5-year DFS 80.3% vs 73.9%. OS (not primary) HR 0.63 (0.43-0.94); 8-year OS 90.3% vs 83.2%. Musculoskeletal adverse events 18.5% vs 11.5%.
  14. 2025 · Treatment · TargetedSTELLAR-303: zanzalintinib plus atezolizumab versus regorafenibIn people whose bowel cancer kept growing after standard treatment and was not MSI-high, zanzalintinib plus atezolizumab was compared with regorafenib. Median survival was about eleven months with the combination and about nine months with regorafenib. Severe treatment-related side effects were more common with the combination.Open-label phase 3, 901 patients with previously treated metastatic colorectal cancer without MSI-H or dMMR tumours, randomised 1:1 to zanzalintinib-atezolizumab (n=451) or regorafenib (n=450). Median OS 10.9 vs 9.4 months (stratified HR 0.80, 95% CI 0.69-0.93; P=0.0045; ITT; data cutoff April 2025). Interim OS in patients without liver metastases: HR 0.79 (0.61-1.03; P=0.087). Grade 3 or worse treatment-related adverse events 60% vs 37%.
  15. 2025 · Detection · Detection & preventionCOLONPREV: colonoscopy versus FIT invitation, 10-year cancer mortalityIn Spain, about 57,000 people aged 50 to 69 were randomly invited to either one colonoscopy or a stool blood test every two years. Fewer took part in screening after a colonoscopy invitation. After ten years, deaths from colorectal cancer were similar in the two groups, about two in a thousand. The trial had no unscreened group.COLONPREV: pragmatic randomised non-inferiority trial at 15 Spanish hospitals, ages 50-69; 57,404 randomised 1:1 to an invitation for one-time colonoscopy or biennial FIT. Participation in any form of screening 31.8% vs 39.9% (RR 0.79, 95% CI 0.77-0.82). Colorectal cancer mortality at 10 years 0.22% (55 deaths) vs 0.24% (60): risk difference -0.02 (95% CI -0.10 to 0.06), RR 0.92 (0.64-1.32), colonoscopy relative to FIT; non-inferiority P=0.0005 against a margin of 0.16 percentage points.
  16. 2026 · Treatment · ImmunotherapyATOMIC: atezolizumab plus adjuvant FOLFOX in dMMR stage III colonIn ATOMIC, people with stage III colon cancer with mismatch repair deficiency were randomly given FOLFOX alone or with atezolizumab, an immunotherapy continued for a year. At three years about eighty-six in a hundred were alive without recurrence with atezolizumab, against about seventy-six. Severe side effects were more common with it. The FDA approved the combination on 8 October 2026.ATOMIC (Sinicrope 2026): phase 3, 712 patients with resected stage III dMMR colon cancer randomised 1:1 to atezolizumab plus mFOLFOX6 for 6 months then atezolizumab (12 months in total; n=355) or mFOLFOX6 alone (n=357). Median follow-up 40.9 months: 3-year DFS 86.3% (95% CI 81.8-89.8) vs 76.2% (70.9-80.6); HR 0.50 (0.35-0.73), P<0.001. Grade 3-4 adverse events 84.1% vs 71.9%. 5-year OS 89.7% vs 87.9% (HR 0.90, 0.55-1.47), immature.
  17. 2026 · Treatment · Chemotherapy⚑ needs reviewCIRCULATE (AIO/ABCSG): chemotherapy for ctDNA-positive stage II colonIn CIRCULATE, people with stage II colon cancer had a blood test for tumour DNA after surgery. Those with a positive test were randomly assigned to chemotherapy or observation. The trial closed early when funding ended, with only twenty-six people in the chemotherapy group, and its main comparison did not reach its goal.CIRCULATE (AIO/ABCSG): stage II, MMR-proficient colon cancer; ctDNA-positive patients randomised 2:1 to chemotherapy (capecitabine with or without oxaliplatin; n=26) or observation. Closed early (funding). Primary ITT endpoint, DFS, not met: 3-year DFS 61% vs 38%, HR 0.55 (95% CI 0.21 to 1.48), one-sided P=0.12. Per-protocol: 77% vs 38%, HR 0.31 (0.09 to 1.03). 3-year DFS ctDNA-positive vs ctDNA-negative: 52% vs 87%, HR 4.28.
  18. 2026 · Treatment · ChemotherapyCOMMIT: mFOLFOX6, bevacizumab and atezolizumab vs atezolizumabIn COMMIT, people with advanced bowel cancer with mismatch repair deficiency, not yet treated, were randomly assigned to atezolizumab, an immunotherapy, alone or with chemotherapy and bevacizumab. The trial closed early. Time before progression or death was longer with the combination, about two years against about five months at the median, but severe side effects were nearly twice as common.COMMIT (NRG-GI004/SWOG-S1610): open-label phase 3, first-line dMMR/MSI-H metastatic colorectal cancer; mFOLFOX6, bevacizumab and atezolizumab (n=41) versus atezolizumab alone (n=41); a third arm closed (n=20). ITT PFS, interim: HR 0.439 (95% CI 0.23 to 0.84), P=0.0103. November 2025 cutoff: HR 0.42 (0.22 to 0.80); median PFS 24.5 vs 5.3 months. ORR 86.1% vs 46.0%. Grade 3 or worse adverse events 82.9% vs 43.9%. OS HR 1.04 (0.47 to 2.28).

Then and now

  • 5-year relative survival · United States, SEER 8 registries

    • 49.41%

      1975 (observed value)

    • 62.3%

      1990 (observed value)

    • 67.73%

      2018 (observed value; last observed year in the table)

    Colon and rectum combined, all races, both sexes; the SEER 8 five-year relative survival series (1975-2018) on the SEER Cancer Stat Facts colorectal page. The page states no age range or stage mix for this series. SEER has no colon-only Cancer Stat Facts page: its index lists 'Colon and Rectum' as one site (checked 22 October 2026).

    Relative survival excludes deaths from other causes. 'Observed' is the table's label for the yearly value, as opposed to the Joinpoint modeled-trend value in the adjacent column. The column is headed 'Year' and the chart note says only '1975-2018'; the page does not say the year is the year of diagnosis. Survival measured from diagnosis can rise through earlier detection (lead-time bias) as well as through treatment, and the page does not say how much of the change in this series is due to screening, stage at diagnosis, treatment or other factors, so read it beside the mortality values. Observed values are used because the 2019-2023 entries in the table are modeled-trend values only. Compare only with other values from this same SEER 8 series. The observed value rose from 49.41% (1975) to 62.30% (1990), was highest in 2016 (69.04%) and was 67.73% in 2018; the modeled-trend values for 1975, 1990 and 2018 are 48.63%, 62.59% and 68.04%.

    Source

  • Age-standardised mortality · United States

    • 28.08per 100,000

      1975 (observed)

    • 24.71per 100,000

      1990 (observed)

    • 13.39per 100,000

      2018 (observed)

    • 12.57per 100,000

      2024 (observed; last year in the table)

    Colon and rectum combined, all races, both sexes; age-adjusted death rate from US mortality data (not SEER registries), the 'Death Rate - U.S.' column of the table on the SEER Cancer Stat Facts colorectal page.

    Death rates describe the whole US population and are not measured from the date of diagnosis, so they are not affected by lead time the way survival from diagnosis is; the page attributes the change to no cause (incidence, screening, treatment and population ageing all act on this rate). The highest observed value in the 1975-2024 table is 28.59 in 1976 (28.56 in 1978). The page headline gives 12.7 per 100,000 for 2020-2024 deaths and says age-adjusted death rates fell on average 1.3% a year over 2015-2024; use the headline or the yearly values, not both. Screened cancers are best read with mortality and survival side by side.

    Source

  • 5-year relative survival · United States, SEER 21 (excluding IL)

    • 65.4%

      2016-2022, all stages

    • 91.3%

      2016-2022, localized stage (34% of cases)

    • 75.2%

      2016-2022, regional stage (37% of cases)

    • 16.9%

      2016-2022, distant stage (23% of cases)

    Colon and rectum combined, all races, both sexes; the headline figure and the stage table on the SEER Cancer Stat Facts colorectal page (SEER 21, excluding IL; stage by SEER Combined Summary Stage). The page states no age range beside them.

    A different registry set and time basis from the SEER 8 series, so these values are not a continuation of the 1975-2018 values and must not be subtracted from them. The page labels the period '2016-2022' and does not say whether it is the diagnosis period. The stage mix is 34% localized, 37% regional, 23% distant and 6% unstaged (49.7% five-year survival for unstaged); the page's text separately says 33.8% are diagnosed at the local stage and that stage at diagnosis has a strong influence on survival. Relative survival excludes deaths from other causes; the page does not discuss lead-time bias or overdiagnosis and gives no breakdown by biomarker (for example mismatch repair status or BRAF V600E) or by colon versus rectum.

    Source

  • 5-year relative survival · United States (SEER-based, via American Cancer Society)

    • 63%

      Diagnosed 2014-2020, all SEER stages combined

    • 91%

      Diagnosed 2014-2020, localized stage

    • 74%

      Diagnosed 2014-2020, regional stage

    • 13%

      Diagnosed 2014-2020, distant stage

    People diagnosed with cancer of the colon (rectum excluded) between 2014 and 2020, by SEER stage, as given on the American Cancer Society page 'Survival Rates for Colorectal Cancer' (last revised 13 January 2026). The page says the figures come from the SEER database and does not say which SEER registry set, age range or sexes they cover.

    Secondary source: the American Cancer Society summarises SEER data, and the registry set and survival-measure details behind these figures are not stated on its page, so they were not checked against SEER itself. This is the only colon-only series on the page. It covers a different diagnosis period (2014-2020) and comes from a different source than the SEER 21 values, so do not subtract or compare the two directly. The ACS page says the numbers apply to stage at first diagnosis, are grouped as SEER localized, regional and distant (not AJCC stage), and that people diagnosed now may have a better outlook because the figures describe people diagnosed and treated at least 5 years earlier.

    Source

Survival figures are affected by earlier detection (lead time) and overdiagnosis, and by changes in who is diagnosed. For screened cancers, read them alongside mortality.

Evidence chains

A lane follows one defined population and setting, where each new trial tests against what was then standard. Compare only within a lane, and read every comparison as illustrative.

Colon cancer after resection: 3-year disease-free survival

Resected colon cancer given adjuvant treatment. The fluorouracil trials listed first enrolled stage B2 and C disease (levamisole), Dukes' B and C colon cancer (IMPACT) and mostly stage II colon or rectal cancer (QUASAR); X-ACT enrolled stage III. MOSAIC and NSABP C-07 enrolled stage II and III, the bevacizumab and cetuximab trials stage II and III or stage III, IDEA stage III, and ATOMIC only stage III with mismatch repair deficiency. · Adjuvant · endpoint: DFS

Results are not comparable across trials.

Populations, eras, control regimens and what is being tested differ. The 1990 levamisole trial used older staging and reported no disease-free survival value here, so it is listed but not drawn. IMPACT reported 3-year event-free survival, a different endpoint, against a control group that its abstract does not describe, and QUASAR reported all-cause mortality, so both are listed but not drawn. X-ACT compared oral capecitabine with bolus fluorouracil and leucovorin and is recorded here with hazard ratios and no 3-year rates, so it is listed but not drawn. MOSAIC and NSABP C-07 tested adding oxaliplatin; the bevacizumab and cetuximab trials tested adding an antibody to oxaliplatin-based chemotherapy (the cetuximab figure is the KRAS wild-type primary analysis). IDEA is a pooled comparison of 3 against 6 months of treatment, so the 3-month arm is drawn as the experimental arm and its non-inferiority margin was not met overall; the component IDEA trials used other margins and endpoints and are not plotted. ATOMIC enrolled only dMMR tumours and its control is mFOLFOX6. Comparisons are only within each trial's own arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

020406080100DFS at 36 months (%)200420042010201020182026MOSAICNSABP C-07NSABP C-08N0147IDEAATOMIC
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
DFS at 36 months, rate, across 6 trials in this lane. MOSAIC (2004): experimental 78.2 % vs control 72.9 %; NSABP C-07 (2004): experimental 76.1 % vs control 71.8 %; NSABP C-08 (2010): experimental 77.4 % vs control 75.5 %; N0147 (2010): experimental 71.5 % vs control 74.6 %; IDEA (2018): experimental 74.6 % vs control 75.5 %; ATOMIC (2026): experimental 86.3 % vs control 76.2 %.
Data table
TrialYearEndpointControl (%)Experimental (%)
MOSAIC2004DFS72.9 %78.2 %
NSABP C-072004DFS71.8 %76.1 %
NSABP C-082010DFS75.5 %77.4 %
N01472010DFS74.6 %71.5 %
IDEA2018DFS75.5 %74.6 %
ATOMIC2026DFS76.2 %86.3 %

Metastatic, first-line chemotherapy: overall survival

Previously untreated metastatic or advanced colorectal cancer, in randomised trials that added irinotecan, oxaliplatin or bevacizumab to first-line chemotherapy (reported 2000 to 2004) or added oxaliplatin to an irinotecan-based regimen with or without bevacizumab (GONO 2007, TRIBE 2014). GONO and TRIBE enrolled adults up to age 75. · Metastatic · endpoint: OS

Results are not comparable across trials.

The control arms differ: bolus fluorouracil and leucovorin (Saltz), infusional fluorouracil and folinate (Douillard, de Gramont), chronomodulated fluorouracil and leucovorin (Giacchetti), irinotecan-based IFL plus placebo (Hurwitz), FOLFIRI (GONO) and FOLFIRI plus bevacizumab (TRIBE). The added drug differs as well (irinotecan, oxaliplatin or bevacizumab; oxaliplatin added to an irinotecan regimen in GONO and TRIBE), as do fluorouracil schedules, follow-up and later treatments. The Hurwitz, GONO and TRIBE controls already contain irinotecan, so their medians are not on the same footing as the fluorouracil-only controls. Overall survival was a secondary endpoint in GONO and TRIBE, whose primary endpoints were response rate and progression-free survival. GONO and TRIBE are one milestone dated 2007 (GONO reported in 2007, TRIBE in 2014), so both are placed at 2007 on the chart. TRIBE2 reported PFS2 and has no overall survival result, so it has no entry. Medians are drawn as published; compare only each trial's own two arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

010203040OS, median (months)2000200020002000200420072007douillard-irinotecan-5fu-lvirinotecan-study-group-iflgiacchetti-oxaliplatin-chronomodulateddegramont-lv5fu2-oxaliplatinhurwitz-ifl-bevacizumabgono-folfoxiriTRIBE
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
OS, median, across 7 trials in this lane. douillard-irinotecan-5fu-lv (2000): experimental 17.4 months vs control 14.1 months; irinotecan-study-group-ifl (2000): experimental 14.8 months vs control 12.6 months; giacchetti-oxaliplatin-chronomodulated (2000): experimental 19.4 months vs control 19.9 months; degramont-lv5fu2-oxaliplatin (2000): experimental 16.2 months vs control 14.7 months; hurwitz-ifl-bevacizumab (2004): experimental 20.3 months vs control 15.6 months; gono-folfoxiri (2007): experimental 22.6 months vs control 16.7 months; TRIBE (2007): experimental 31 months vs control 25.8 months.
Data table
TrialYearEndpointControl (months)Experimental (months)
douillard-irinotecan-5fu-lv2000OS14.1 months17.4 months
irinotecan-study-group-ifl2000OS12.6 months14.8 months
giacchetti-oxaliplatin-chronomodulated2000OS19.9 months19.4 months
degramont-lv5fu2-oxaliplatin2000OS14.7 months16.2 months
hurwitz-ifl-bevacizumab2004OS15.6 months20.3 months
gono-folfoxiri2007OS16.7 months22.6 months
TRIBE2007OS25.8 months31 months

Metastatic, first line, wild-type tumours: EGFR antibody versus bevacizumab, overall survival

Previously untreated metastatic colorectal cancer without KRAS or RAS mutations, in head-to-head trials of an anti-EGFR antibody (cetuximab or panitumumab) against bevacizumab, each added to chemotherapy. FIRE-3 enrolled KRAS exon 2 wild-type tumours, CALGB/SWOG 80405 KRAS wild-type tumours and PARADIGM, a Japanese trial, RAS wild-type tumours. · Metastatic · endpoint: OS

Results are not comparable across trials.

Every entry compares an anti-EGFR antibody (experimental arm) with bevacizumab (control arm), each with chemotherapy. PRIME is not an entry because its control was chemotherapy alone. The populations differ in how deeply tumours were tested (KRAS exon 2 in FIRE-3, KRAS in 80405, RAS in PARADIGM), as do the chemotherapy backbones (FOLFIRI in FIRE-3, mFOLFOX6 or FOLFIRI in 80405, mFOLFOX6 in PARADIGM) and the primary endpoints: objective response in FIRE-3, so its survival result is not primary, and overall survival in 80405 and PARADIGM. PARADIGM's survival analyses use the as-treated population (802 of 823 randomised participants), not all randomised participants. Its protocol was revised in May 2019, after enrolment ended and before data analysis, to make survival in left-sided tumours the primary outcome; the left-sided and the overall result are both drawn, as two PARADIGM markers at 2017 with the left-sided result listed first. The PARADIGM survival curves were not parallel and appeared to cross at about 28 months, so a median or hazard ratio may not summarise the difference. Not drawn because they are exploratory or retrospective: PARADIGM's right-sided results, the pooled analysis of six trials by tumour side, and the later post hoc FIRE-3 RAS wild-type subgroup. Compare only each trial's own two arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

010203040OS, median (months)2010201720172017FIRE-3CALGB/SWOG 80405PARADIGMPARADIGM
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
OS, median, across 4 trials in this lane. FIRE-3 (2010): experimental 28.7 months vs control 25 months; CALGB/SWOG 80405 (2017): experimental 30 months vs control 29 months; PARADIGM (2017): experimental 37.9 months vs control 34.3 months; PARADIGM (2017): experimental 36.2 months vs control 31.3 months.
Data table
TrialYearEndpointControl (months)Experimental (months)
FIRE-32010OS25 months28.7 months
CALGB/SWOG 804052017OS29 months30 months
PARADIGM2017OS34.3 months37.9 months
PARADIGM2017OS31.3 months36.2 months

Metastatic, after standard chemotherapy: overall survival

Metastatic colorectal cancer that progressed on standard treatment. The control arm in each plotted trial is best supportive care alone or with placebo. · Metastatic · endpoint: OS

Results are not comparable across trials.

Prior-therapy requirements, number of earlier lines, region, biomarker selection and crossover differ. CO.17 enrolled people with detectable EGFR before KRAS testing was used to select treatment, and its control was best supportive care alone; CORRECT, RECOURSE, FRESCO and FRESCO-2 used placebo. FRESCO was run in China and FRESCO-2 in a global population. SUNLIGHT compared trifluridine-tipiracil with and without bevacizumab, and STELLAR-303 compared zanzalintinib with atezolizumab against regorafenib; both have an active control that does not fit with the placebo-controlled trials, so SUNLIGHT is not plotted and STELLAR-303 is not in this lane. The panitumumab trial against best supportive care has no median overall survival values in this record and is not drawn. BEACON, in BRAF V600E-mutated disease with an active control, is a different population and is not in this lane. Compare only each trial's own two arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

02468101214OS, median (months)20042012201520182018CO.17CORRECTRECOURSEFRESCOFRESCO-2
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
OS, median, across 5 trials in this lane. CO.17 (2004): experimental 6.1 months vs control 4.6 months; CORRECT (2012): experimental 6.4 months vs control 5 months; RECOURSE (2015): experimental 7.1 months vs control 5.3 months; FRESCO (2018): experimental 9.3 months vs control 6.6 months; FRESCO-2 (2018): experimental 7.4 months vs control 4.8 months.
Data table
TrialYearEndpointControl (months)Experimental (months)
CO.172004OS4.6 months6.1 months
CORRECT2012OS5 months6.4 months
RECOURSE2015OS5.3 months7.1 months
FRESCO2018OS6.6 months9.3 months
FRESCO-22018OS4.8 months7.4 months

MSI-H or dMMR metastatic, first line: progression-free survival

Previously untreated unresectable or metastatic colorectal cancer with MSI-H or dMMR status. · Metastatic · endpoint: PFS

Results are not comparable across trials.

Both trials compared an immunotherapy with investigator's-choice chemotherapy, but they are parallel experiments, not steps in a sequence. KEYNOTE-177 reports a median PFS; CheckMate 8HW reports a 24-month PFS rate for its first-line comparison, so it is listed but not drawn on the same axis. MSI-H or dMMR status was confirmed by different methods, follow-up and crossover differ, and CheckMate 8HW's first-line analysis covers part of its randomised population. COMMIT, a 2026 randomised trial in the same population, reports mFOLFOX6 with bevacizumab and atezolizumab against atezolizumab alone (its chemotherapy-only arm closed early, after 20 patients), so it has no chemotherapy comparison and is not in this lane. The regimens were not compared with each other; read each trial's own two arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

05101520PFS, median (months)2020KEYNOTE-177
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
PFS, median, across 1 trial in this lane. KEYNOTE-177 (2020): experimental 16.5 months vs control 8.2 months.
Data table
TrialYearEndpointControl (months)Experimental (months)
KEYNOTE-1772020PFS8.2 months16.5 months
CheckMate 8HW2024PFS (not plotted: different measure or time point)14 %72 %

Screening trials: colorectal cancer mortality

Asymptomatic adults aged about 50 to 64 invited to screening (faecal occult blood testing, once-only flexible sigmoidoscopy, a single colonoscopy invitation) compared with no invitation or usual care. · Screening · endpoint: Colorectal cancer mortality

Results are not comparable across trials.

The tests, age ranges, uptake and follow-up differ, and these are intention-to-screen comparisons of an invitation, not of the test itself. The faecal occult blood trials and the sigmoidoscopy trials report hazard ratios or relative risks without a risk in each arm, so only NordICC's 10-year risk of death from colorectal cancer can be drawn here. NordICC's per-protocol result is not a two-arm comparison and is not used. COLONPREV compared an invitation to colonoscopy with an invitation to faecal immunochemical testing and has no unscreened arm, so it is not in this lane. Read each trial's own two arms; the values cannot be ranked across tests.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

020406080100Colorectal cancer mortality (%)2022NordICC
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
Colorectal cancer mortality, rate, across 1 trial in this lane. NordICC (2022): experimental 0.28 % vs control 0.31 %.
Data table
TrialYearEndpointControl (%)Experimental (%)
NordICC2022Colorectal cancer mortality0.31 %0.28 %

Standard of care, then and now

How the usual sequence of treatment changed for one setting. Choose a point in time; each step links to the milestone behind it.

Colon cancer after surgery: adjuvant treatment

Resected colon cancer. The stage studied differs by trial, and some steps apply only to selected groups (stage III disease, mismatch repair-deficient tumours, or stage II disease with a circulating tumour DNA result). · Adjuvant

  1. Chemotherapy1Fluorouracil + levamisoleAdjuvant fluorouracil plus levamisole for stage III colon cancer (1990)
  2. Chemotherapy2Fluorouracil + folinic acidAdjuvant fluorouracil with folinic acid: IMPACT and QUASAR (stage II) (1995)
  3. Chemotherapy3Fluorouracil/leucovorin + oxaliplatinMOSAIC: oxaliplatin added to adjuvant fluorouracil for colon cancer (2004)
  4. Chemotherapy4Oral capecitabine instead of FU/LVX-ACT: adjuvant capecitabine for stage III colon cancer (2005)
  5. Chemotherapy53 versus 6 months of oxaliplatin regimensIDEA: 3 versus 6 months of adjuvant oxaliplatin chemotherapy (2018)
  6. Chemotherapy6Stage II: ctDNA-guided chemotherapyDYNAMIC: ctDNA-guided adjuvant therapy in stage II colon cancer (2022)
  7. Immunotherapy7Stage III dMMR: atezolizumab + FOLFOXATOMIC: atezolizumab plus adjuvant FOLFOX in dMMR stage III colon (2026)
Today. Based only on the trials cited in each step. ATOMIC added atezolizumab to FOLFOX in stage III dMMR disease (FDA, 8 October 2026). DYNAMIC used ctDNA to guide stage II chemotherapy; DYNAMIC-III, in stage III, missed its de-escalation margin.
Text version
  • 1995: Fluorouracil + levamisole → Fluorouracil + folinic acid
  • 2005: Fluorouracil + levamisole → Fluorouracil + folinic acid → Fluorouracil/leucovorin + oxaliplatin → Oral capecitabine instead of FU/LV
  • 2018: Fluorouracil + levamisole → Fluorouracil + folinic acid → Fluorouracil/leucovorin + oxaliplatin → Oral capecitabine instead of FU/LV → 3 versus 6 months of oxaliplatin regimens
  • Today: Fluorouracil + levamisole → Fluorouracil + folinic acid → Fluorouracil/leucovorin + oxaliplatin → Oral capecitabine instead of FU/LV → 3 versus 6 months of oxaliplatin regimens → Stage II: ctDNA-guided chemotherapy → Stage III dMMR: atezolizumab + FOLFOX

Metastatic colorectal cancer: first-line options

Previously untreated metastatic colorectal cancer; later options depend on biomarkers (RAS wild-type, MSI-H or dMMR, BRAF V600E) · Metastatic

  1. Chemotherapy1FU/LV + irinotecanIrinotecan with fluorouracil as first-line therapy for metastatic CRC (2000)
  2. Chemotherapy2FU/LV + oxaliplatinOxaliplatin added to infusional fluorouracil in advanced CRC (2000)
  3. Targeted3IFL + bevacizumabBevacizumab added to first-line irinotecan-based chemotherapy (2004)
  4. Targeted4FOLFOXIRI + bevacizumabFOLFOXIRI with bevacizumab: GONO trial, TRIBE and TRIBE2 (2007)
  5. Targeted5RAS wild-type: anti-EGFR + chemotherapyFirst-line anti-EGFR trials and extended RAS testing (PRIME, FIRE-3) (2010)
  6. Immunotherapy6MSI-H/dMMR: pembrolizumabKEYNOTE-177: pembrolizumab first-line for MSI-H/dMMR metastatic CRC (2020)
  7. Immunotherapy7MSI-H/dMMR: nivolumab + ipilimumabCheckMate 8HW: nivolumab plus ipilimumab in MSI-H/dMMR mCRC (2024)
  8. Targeted8BRAF V600E: encorafenib, cetuximab, mFOLFOX6BREAKWATER: first-line encorafenib, cetuximab and chemotherapy (2024)
Today. Based only on the trials cited in each step. Two further first-line trials are listed: nivolumab with ipilimumab for MSI-H or dMMR tumours, and an encorafenib regimen for BRAF V600E-mutated tumours.
Text version
  • 2000: FU/LV + irinotecan → FU/LV + oxaliplatin
  • 2004: FU/LV + irinotecan → FU/LV + oxaliplatin → IFL + bevacizumab
  • 2020: FU/LV + irinotecan → FU/LV + oxaliplatin → IFL + bevacizumab → FOLFOXIRI + bevacizumab → RAS wild-type: anti-EGFR + chemotherapy → MSI-H/dMMR: pembrolizumab
  • Today: FU/LV + irinotecan → FU/LV + oxaliplatin → IFL + bevacizumab → FOLFOXIRI + bevacizumab → RAS wild-type: anti-EGFR + chemotherapy → MSI-H/dMMR: pembrolizumab → MSI-H/dMMR: nivolumab + ipilimumab → BRAF V600E: encorafenib, cetuximab, mFOLFOX6

Reversals and dead ends

Approaches that were abandoned or withdrawn, and trials whose results were later questioned, are part of the story.

  • 2008

    KRAS mutation predicts no benefit: anti-EGFR limited to wild-type

    Cetuximab and panitumumab were first approved for bowel cancers carrying the EGFR receptor. In 2008, two analyses of randomised trials showed that people whose tumours had a KRAS mutation gained no benefit, while those without it did. The FDA later revised the cetuximab label and then limited its use to tumours without the mutation.

    Biomarker analyses of two randomised trials of anti-EGFR antibody plus best supportive care versus best supportive care alone. Panitumumab (KRAS known in 427 of 463): PFS HR 0.45 (95% CI 0.34 to 0.59) in wild-type vs 0.99 (0.73 to 1.36) in mutant tumours. Cetuximab, CO.17 (K-ras tested in 394 of 572): wild-type OS median 9.5 vs 4.8 months, HR 0.55 (0.41 to 0.74); mutant OS HR 0.98 (P=0.89). FDA limited cetuximab to K-ras wild-type disease on 6 July 2012.

  • 2010

    Adjuvant bevacizumab and cetuximab trials in colon cancer

    Four large randomised trials added a targeted drug, bevacizumab or cetuximab, to oxaliplatin-based chemotherapy after surgery for colon cancer. None improved disease-free survival. In one trial the authors noted overall survival data suggesting a possible harm with bevacizumab, and cetuximab added side effects. Both drugs had shown benefit in advanced disease, which is a different setting.

    Adjuvant colon cancer, DFS primary endpoint. NSABP C-08 (bevacizumab + mFOLFOX6; 2672 analysed): HR 0.89 (95% CI 0.76-1.04), P=0.15. AVANT (3451 randomised; stage III primary analysis): HR 1.17 (0.98-1.39) bevacizumab-FOLFOX4 and 1.07 (0.90-1.28) bevacizumab-XELOX vs FOLFOX4; OS HR 1.27 (1.03-1.57) for bevacizumab-FOLFOX4. N0147 (KRAS wild-type, mFOLFOX6 with or without cetuximab): 3-year DFS 71.5% with vs 74.6% without, HR 1.21 (0.98-1.49). PETACC-8 (KRAS exon 2 wild-type, 1602): HR 1.05 (0.85-1.29).

  • 2021

    PRODIGE 7: HIPEC added to cytoreductive surgery (peritoneal CRC)

    PRODIGE 7 randomly assigned 265 people with colorectal cancer that had spread to the lining of the abdomen to surgery that removes all visible tumour, with or without heated chemotherapy washed through the abdomen. Median survival was about forty-one to forty-two months in both groups, and later complications were more frequent with the heated chemotherapy.

    PRODIGE 7 (Quénet 2021): phase 3, 265 patients with colorectal peritoneal metastases (PCI 25 or less) after complete cytoreduction, randomised to surgery plus oxaliplatin HIPEC (n=133) or surgery alone (n=132). Median OS 41.7 months (95% CI 36.2-53.8) vs 41.2 (35.1-49.7); HR 1.00 (95.37% CI 0.63-1.58), P=0.99; median follow-up 63.8 months. Grade 3 or worse adverse events at 60 days: 26% vs 15% (P=0.035).

Today and tomorrow

Standard of care today

Screening

Guideline abstracts and the USPSTF page were read; the guideline full texts were not. The ACS 2018 abstract recommends regular screening for average-risk adults from age 45 with a high-sensitivity stool-based test or a structural examination (a qualified recommendation for starting at 45; the recommendation for regular screening from 50 is strong). The ACS 2026 update abstract reaffirms starting at 45 and continuing to 75 for people with a life expectancy over 10 years; recommends a next-generation multitarget stool DNA test and a multitarget stool RNA test, alongside annual high-sensitivity FIT and guaiac-based FOBT, as preferred stool-based options (the abstract's wording leaves unclear which testing interval applies to which test, so no interval is given here); says blood-based tests, which had lower sensitivity for advanced precancerous lesions and stage I cancers, should be recommended only to people who decline or do not complete preferred tests; and says a positive noncolonoscopy test needs timely colonoscopy, preferably within 6 months. The USPSTF page (final recommendation statement of 18 May 2021) gives grade A for ages 50 to 75, grade B for 45 to 49 and grade C (selective screening) for 76 to 85; its strategy list does not include blood-based tests, and it states that the recommendation does not include serum tests, urine tests or capsule endoscopy. FDA device decisions (PMA database pages read; these are device approvals, not drug approvals):

  • Cologuard (P130017) on 2014-08-11 for adults 50 or older at typical average risk, with supplement S029 (2019-09-20) expanding the indicated age range to 45 or older
  • ColoSense, a stool test for RNA markers and occult haemoglobin (P230001), on 2024-05-03 and Cologuard Plus (P230043) on 2024-10-03, both for adults 45 or older at average risk
  • and two blood-based tests, Shield (P230009
  • 2024-07-26
  • average-risk individuals aged 45 or older
  • supplement S010 of 2026-07-22 approved a change to a methylation-only classification workflow) and SimpleScreen CRC (P250029
  • 2026-07-24
  • adults 45 or older at average risk). Each of the five original approval statements says a positive result should be followed by colonoscopy and that the test is not a replacement for diagnostic colonoscopy in high-risk individuals. Other supplements to these approvals were not read, so current labelling is not described. Two randomised invitation trials of colonoscopy are summarised here: NordICC (see hooks) and COLONPREV, which compares two screening strategies and has no unscreened arm: 57,404 people aged 50 to 69 in Spain were randomised to an invitation for one-time colonoscopy or for biennial FIT
  • 10-year colorectal cancer mortality was 0.22% (55 deaths) with colonoscopy and 0.24% (60 deaths) with FIT, FIT was reported as non-inferior (margin 0.16 percentage points), and participation in any form of screening was 31.8% versus 39.9%
Guideline: Wolf AMD, Fontham ETH, Church TR, Flowers CR, Guerra CE, LaMonte SJ, et al 2018Wolf AMD, Hoffman RM, Walter LC, Zeigler-Johnson C, Church TR, Guerra CE, et al 2026Guideline: Davidson KW, Barry MJ, Mangione CM, Cabana M, Caughey AB, Davis EM, et al 2021US Preventive Services Task Force 2021US Food and Drug Administration 2014US Food and Drug Administration 2019US Food and Drug Administration 2024US Food and Drug Administration 2024US Food and Drug Administration 2024US Food and Drug Administration 2026US Food and Drug Administration 2026Bretthauer M, Løberg M, Wieszczy P, Kalager M, Emilsson L, Garborg K, et al 2022Castells A, Quintero E, Bujanda L, Castán-Cameo S, Cubiella J, Díaz-Tasende J, et al 2025
Adjuvant

Resected colon cancer. FDA letters read in full:

  • oxaliplatin (Eloxatin
  • NDA 21-492/S-004
  • letter signed 2004-11-04) in combination with infusional 5-FU/leucovorin for adjuvant treatment of stage III colon cancer after complete resection of the primary tumour, an indication the letter says is based on an improvement in disease-free survival with no demonstrated benefit in overall survival after a median follow-up of 4 years
  • capecitabine (Xeloda
  • NDA 20-896/S-016
  • signed 2005-06-15) as a single agent for adjuvant treatment of Dukes' C colon cancer after complete resection 'when treatment with fluoropyrimidine therapy alone is preferred' (the letter states that capecitabine was non-inferior to 5-FU/leucovorin for disease-free survival, that neither capecitabine nor combination chemotherapy prolongs overall survival, and that combination chemotherapy improved disease-free survival compared with 5-FU/leucovorin)
  • and atezolizumab (Tecentriq
  • BLA 761034/S-065
  • e-signed 2026-10-08) in combination with a fluoropyrimidine and oxaliplatin for adjuvant treatment of adult and paediatric patients 2 years of age and older with stage III mismatch repair-deficient (dMMR) colon cancer. The atezolizumab letter contains no accelerated-approval wording, so this is recorded as a regular approval (an inference from the letter)
  • it lists postmarketing commitment 5080-1, a validation study for a companion diagnostic for dMMR stage 3 colon cancer, with the final report due 09/2029. In the EU, the Xeloda product page lists adjuvant treatment after surgery of stage III (Dukes' stage C) colon cancer (variation II/0009: opinion 17 February 2005, Commission decision 30 March 2005)
  • the EU position on atezolizumab for this indication was not checked. The ESMO Express Update on physical exercise in localised colon cancer (ESMO Open 2026;11(2):106019
  • full text read) says that patients with resected stage III or high-risk stage II colon cancer should be informed of the available evidence suggesting survival benefits from a structured exercise programme [I, A] and that, after discussion and joint decision-making, participation of eligible patients in a structured exercise programme can be recommended [II, B
  • ESMO-MCBS v2.0 score A]
  • it also gives adjuvant treatment algorithms for stage II and stage III disease, which are not described here. The ESMO localised colon cancer guideline (2020) and NCCN Colon Cancer (v2.2021 and v3.2024
  • the latest NCCN colon article found in PubMed, the live version was not retrieved) are cited by identifier only and their text was not read
U 2004U 2005U 2026European Medicines Agency 2026European Medicines Agency 2005Guideline: Pentheroudakis G, Argilés G, Arnold D, Smyth E, Ducreux M 2026Guideline: Argilés G, Tabernero J, Labianca R, Hochhauser D, Salazar R, Iveson T, et al 2020Guideline: Benson AB, Venook AP, Al-Hawary MM, Arain MA, Chen YJ, Ciombor KK, et al 2021Guideline: Benson AB, Venook AP, Adam M, Chang G, Chen YJ, Ciombor KK, et al 2024
Neoadjuvant

Locally advanced colon cancer before surgery. FOxTROT (Morton 2023; abstract read) randomised patients with radiologically staged T3-4 N0-2 M0 colon cancer 2:1 to 6 weeks of oxaliplatin-fluoropyrimidine chemotherapy before surgery plus 18 weeks after (699 patients) or 24 weeks after surgery (354 patients); patients with RAS-wildtype tumours could also be randomised to panitumumab during neoadjuvant treatment, which the abstract says did not enhance the benefit. Residual disease or recurrence within 2 years was 16.9% versus 21.5% (rate ratio 0.72; 95% CI 0.54 to 0.98; P=0.037) and resection was histopathologically complete in 94% versus 89%; the abstract says little benefit was seen in mismatch repair-deficient tumours and reports no survival result. NICHE-2 (Chalabi 2024; abstract read) was a single-arm phase 2 study of neoadjuvant nivolumab plus ipilimumab in 115 people with untreated, locally advanced dMMR colon cancer:

  • 113 (98%) had timely surgery
  • among 111 in the efficacy analysis, pathological response was seen in 109 (98%), major pathological response in 105 (95%) and pathological complete response in 75 (68%)
  • grade 3 or 4 immune-related adverse events occurred in 5 (4%)
  • at a median follow-up of 26 months no patient had a recurrence, and the abstract does not report 3-year disease-free survival, one of the two primary endpoints. No neoadjuvant colon cancer indication appears in the FDA letters or EMA product pages read for this page. The ESMO localised colon cancer guideline (2020) is cited by identifier only and its text was not read. Perioperative dostarlimab in dMMR colon cancer is being tested in AZUR-2 (see horizon)
Morton D, Seymour M, Magill L, Handley K, Glasbey J, Glimelius B, et al 2023Chalabi M, Verschoor YL, Tan PB, Balduzzi S, Van Lent AU, Grootscholten C, et al 2024Guideline: Argilés G, Tabernero J, Labianca R, Hochhauser D, Salazar R, Iveson T, et al 2020ClinicalTrials 2026
Metastatic

Biomarker-defined first-line treatment. FDA letters read in full:

  • pembrolizumab (Keytruda
  • BLA 125514/S-084
  • e-signed 2020-06-29) for first-line treatment of unresectable or metastatic MSI-H or dMMR colorectal cancer, with no accelerated-approval wording (recorded as a regular approval, an inference) and a postmarketing commitment (3892-1) for the final KEYNOTE-177 overall-survival analyses
  • nivolumab (Opdivo
  • BLA 125554/S-132
  • e-signed 2025-04-08), which updates the labelled indications to nivolumab with ipilimumab for adult and paediatric patients 12 years and older with unresectable or metastatic MSI-H or dMMR colorectal cancer and single-agent nivolumab for MSI-H or dMMR metastatic disease that progressed after a fluoropyrimidine, oxaliplatin and irinotecan, and whose letter refers to the conversion from accelerated approval to full approval and carries the heading 'SUBPART E FULFILLED' for supplements S-034 and S-063
  • encorafenib (Braftovi
  • NDA 210496) with cetuximab and mFOLFOX6 for BRAF V600E-mutated metastatic colorectal cancer under accelerated approval (S-017, e-signed 2024-12-20), followed by a letter (S-019, e-signed 2026-02-24) covering encorafenib with cetuximab and fluorouracil-based chemotherapy under the heading 'SUBPART H FULFILLED' (recorded as a conversion
  • the letter does not use the word 'converted')
  • and panitumumab (Vectibix
  • BLA 125147) with FOLFOX for first-line treatment of wild-type KRAS (exon 2) metastatic colorectal cancer (S-108, 2014-05-23), with the indicated population changed to wild-type RAS (KRAS and NRAS) on 2017-06-29 (S-207)
  • neither of those two letters contains accelerated-approval wording (recorded as regular, an inference). A Vectibix letter of the same date as S-108 (S-186, 2014-05-23), for monotherapy in wild-type KRAS (exon 2) disease after progression on fluoropyrimidine-, oxaliplatin- and irinotecan-containing regimens, carries the heading 'SUBPART E FULFILLED' and does not use the word 'converted'. EMA pages read (2026-10-22): Keytruda monotherapy for MSI-H or dMMR colorectal cancer in first-line metastatic disease and after previous fluoropyrimidine-based combination therapy
  • Opdivo with ipilimumab for dMMR or MSI-H colorectal cancer as first-line treatment of unresectable or metastatic disease and after prior fluoropyrimidine-based combination chemotherapy
  • and, for BRAF V600E-mutant metastatic disease, Braftovi with cetuximab after prior systemic therapy and with cetuximab and FOLFOX as first-line treatment, and Erbitux with encorafenib after prior systemic therapy and with encorafenib and FOLFOX as first-line treatment (the Braftovi first-line indication and both new Erbitux indications followed CHMP positive opinions of 21 May 2026
  • EMA procedural-steps documents record Commission decisions issued on 19 June 2026 for Braftovi and 25 June 2026 for Erbitux, and the decisions themselves were not read). The ESMO metastatic colorectal cancer guideline is cited by identifier only (2023 edition and a 2026 edition, Ann Oncol 2026;37(6)
  • whether the 2026 text supersedes the 2023 one was not established) and neither text was read. Approvals for chemotherapy backbones and for bevacizumab and cetuximab in earlier lines are covered in the metastatic milestones and were not re-read for this page
US Food and Drug Administration 2020US Food and Drug Administration 2025US Food and Drug Administration 2024US Food and Drug Administration 2026US Food and Drug Administration 2014US Food and Drug Administration 2014US Food and Drug Administration 2017European Medicines Agency 2026European Medicines Agency 2026European Medicines Agency 2026European Medicines Agency, Committee for Medicinal Products for Human Use 2026European Medicines Agency 2026European Medicines Agency 2026European Medicines Agency, Committee for Medicinal Products for Human Use 2026European Medicines Agency 2026Guideline: Cervantes A, Adam R, Roselló S, Arnold D, Normanno N, Taïeb J, et al 2023Guideline: Cremolini C, Chalabi M, Elez E, Fassan M, Gelli M, Goéré D, et al 2026
Relapsed / refractory

Later-line metastatic disease. FDA letters read in full:

  • regorafenib (Stivarga
  • NDA 203085
  • 2012-09-27) for metastatic colorectal cancer previously treated with fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy and, if KRAS wild type, an anti-EGFR therapy
  • trifluridine/tipiracil (Lonsurf
  • NDA 207981
  • 2015-09-22) for the same prior-treatment description (anti-EGFR therapy if RAS wild type), with supplement S-012 (2023-08-02) adding use as a single agent or in combination with bevacizumab
  • and fruquintinib (Fruzaqla
  • NDA 217564
  • 2023-11-08) after fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy, an anti-VEGF therapy and, if RAS wild type and medically appropriate, an anti-EGFR therapy. Biomarker-directed options: encorafenib with cetuximab for BRAF V600E-mutated disease after prior therapy (Braftovi S-006, 2020-04-08
  • no accelerated-approval wording, recorded as regular, an inference)
  • tucatinib with trastuzumab for RAS wild-type, HER2-positive disease that progressed after fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapy (Tukysa S-004, 2023-01-19, accelerated approval)
  • sotorasib with panitumumab for KRAS G12C-mutated disease after the same chemotherapies (Lumakras S-009, 2025-01-16
  • no accelerated-approval wording, recorded as regular, an inference)
  • and trastuzumab deruxtecan for unresectable or metastatic HER2-positive (IHC 3+) solid tumours after prior systemic treatment with no satisfactory alternative (Enhertu S-028, 2024-04-05, accelerated approval
  • not specific to colorectal cancer). Adagrasib with cetuximab for KRAS G12C-mutated disease received accelerated approval on 2024-06-21 (Krazati S-005)
  • a later letter (NDA 216340/S-011, e-signed 2026-09-01) approves the sponsor's voluntary withdrawal of that indication, states that the confirmatory study (Study 849-010) fulfilled the postmarketing requirement but 'did not confirm clinical benefit', and notes that an efficacy supplement would be needed to resume marketing, so the adagrasib colorectal indication is no longer an approved use. For tucatinib, the only later efficacy supplement in the openFDA Drugs@FDA list (queried 2026-10-22) is S-010 (2026-10-07), whose letter covers HER2-positive breast cancer
  • for trastuzumab deruxtecan the later efficacy supplements in the same list were read: S-032 and S-035 (2025-01-27) and S-038 and S-042 (2025-12-15) cover breast cancer, S-039 (2026-04-27) updates labelling from a lung cancer study, and S-041 and S-043 (2026-05-15) cover breast cancer
  • none mentions colorectal cancer or solid tumours. Whether the tucatinib and trastuzumab deruxtecan accelerated approvals have been converted was not established. EMA product pages read: Fruzaqla (authorised 20 June 2024) as monotherapy after standard therapies in people who progressed on or are intolerant to trifluridine-tipiracil or regorafenib
  • Lonsurf (authorised 25 April 2016) with bevacizumab after two prior regimens, and as monotherapy after available therapies. Zanzalintinib with atezolizumab: a Form 8-K filed by Exelixis dated 2026-09-10 (company filing) states that the FDA extended the review period for the NDA (zanzalintinib with atezolizumab, metastatic colorectal cancer) by three months after a submission the FDA deemed a major amendment, with a new action date of March 3, 2027
  • an OncLive report of 2026-09-11 (secondary source) gives the earlier date as December 3, 2026. This is an application under review, not an approval. The ESMO metastatic colorectal cancer guideline is cited by identifier only (2023 edition and a 2026 edition
  • neither text was read)
US Food and Drug Administration 2012US Food and Drug Administration 2015US Food and Drug Administration 2023US Food and Drug Administration 2023US Food and Drug Administration 2020US Food and Drug Administration 2023US Food and Drug Administration 2026US Food and Drug Administration 2025US Food and Drug Administration 2024US Food and Drug Administration 2025US Food and Drug Administration 2025US Food and Drug Administration 2026US Food and Drug Administration 2026US Food and Drug Administration 2024US Food and Drug Administration 2026European Medicines Agency 2026European Medicines Agency 2024Exelixis, Inc 2026OncLive 2026Guideline: Cervantes A, Adam R, Roselló S, Arnold D, Normanno N, Taïeb J, et al 2023Guideline: Cremolini C, Chalabi M, Elez E, Fassan M, Gelli M, Goéré D, et al 2026

Trials to watch

  • AZUR-2 NCT05855200

    In untreated T4N0 or stage III dMMR/MSI-H resectable colon cancer, does perioperative dostarlimab monotherapy differ from standard of care (FOLFOX or CAPOX, or observation after surgery) in event-free survival assessed by blinded independent central review (the registry's primary outcome)? Registry record: randomised open-label phase 3, 892 estimated, GlaxoSmithKline sponsor. Registry status: recruiting; no results posted.

  • CIRCULATE (PRODIGE 70) NCT04120701

    In resected stage II colon cancer (the registry's eligibility also admits high rectal tumours; ages 18 to 75), does a circulating tumour DNA-based decision to give adjuvant chemotherapy change 3-year disease-free survival in people with detectable ctDNA (the registry's primary outcome)? Registry record: randomised phase 3, 1,980 estimated, sponsor Centre Hospitalier Universitaire Dijon; the registry lists one chemotherapy arm and two follow-up arms. Registry status: recruiting; no results posted.

  • MOUNTAINEER-03 NCT05253651

    In previously untreated HER2-positive metastatic colorectal cancer, does tucatinib with trastuzumab and mFOLFOX6 differ from mFOLFOX6 with or without cetuximab or bevacizumab in progression-free survival by blinded independent central review (the registry's primary outcome)? Registry record: randomised open-label phase 3, 400 estimated, sponsor Seagen (a wholly owned subsidiary of Pfizer). Registry status: recruiting; no results posted.

  • CodeBreaK 301 NCT06252649

    In previously untreated KRAS p.G12C-mutant metastatic colorectal cancer, does sotorasib with panitumumab and FOLFIRI differ from FOLFIRI with or without bevacizumab-awwb in progression-free survival per RECIST v1.1 (the registry's primary outcome)? Registry record: randomised open-label active-controlled phase 3, 450 estimated, sponsor Amgen. Registry status: recruiting; no results posted.

  • KANDLELIT-012 (MK-1084-012) NCT06997497

    In first-line KRAS G12C-mutant, locally advanced unresectable or metastatic colorectal cancer, how does calderasib (MK-1084) with cetuximab and mFOLFOX6 compare with mFOLFOX6 with or without bevacizumab? The registry lists dose-limiting toxicity, progression-free survival and, for Part 1, adverse-event and treatment-discontinuation outcomes as primary outcomes. Registry record: randomised phase 3 (open-label in the official title), 477 estimated, sponsor Merck Sharp & Dohme LLC. Registry status: recruiting; no results posted.

  • OrigAMI-2 NCT06662786

    In first-line KRAS/NRAS and BRAF wild-type, unresectable or metastatic left-sided colorectal cancer, does amivantamab with mFOLFOX6 or FOLFIRI differ from cetuximab with mFOLFOX6 or FOLFIRI in progression-free survival assessed by blinded independent central review (the registry's primary outcome)? Registry record: randomised open-label phase 3, 1,000 estimated, sponsor Janssen Research & Development. Registry status: recruiting; no results posted.

References

Show all 261 references

Loading references… Open the list as a page

Every figure was checked against its primary source to the extent possible. Items that could not be fully confirmed are marked “needs review”. Educational history, not medical advice.

Report an error

Links open the primary source. A flag marked ⚑ needs review means a value or claim has not yet been independently confirmed.

2010TreatmentTargeted

majorMetastatic

First-line anti-EGFR trials and extended RAS testing (PRIME, FIRE-3)

In people with untreated bowel cancer that had spread, adding an antibody against the EGF receptor to chemotherapy was linked to better outcomes when the tumour had no RAS gene changes, not when it had them. In one trial's RAS analysis, median survival with panitumumab was about six months longer. Testing widened from KRAS exon 2 to KRAS and NRAS.

PRIME (n=1,183, FOLFOX4 with or without panitumumab): KRAS wild-type median PFS 9.6 vs 8.0 months (HR 0.80, 95% CI 0.66-0.97); prospective-retrospective RAS wild-type analysis (512 patients) median OS 26.0 vs 20.2 months (HR 0.78, 0.62-0.99). Other RAS mutations (17% of KRAS exon 2 wild-type tumours) were associated with inferior outcomes with panitumumab. FIRE-3 (FOLFIRI, KRAS exon 2 wild-type), cetuximab vs bevacizumab: response 62.0% vs 58.0% (primary endpoint, P=0.18); median OS 28.7 vs 25.0 months (HR 0.77, not primary).

Standard before

The 2013 PRIME report starts from the position that KRAS exon 2 mutations predict no benefit from anti-EGFR therapy; panitumumab was approved as monotherapy for chemotherapy-refractory metastatic colorectal cancer, and first-line use with chemotherapy was being tested (2010 PRIME report).

What changed

FDA labelling of panitumumab with FOLFOX as first-line treatment: wild-type KRAS (exon 2) metastatic colorectal cancer (letter of 23 May 2014), later modified to wild-type RAS (KRAS and NRAS) as determined by an FDA-approved test, for first-line use with FOLFOX and for monotherapy after progression (letter of 29 June 2017). A supplement approved on the same day as the first-line indication (BLA 125147/186) added panitumumab monotherapy for wild-type KRAS disease after progression on fluoropyrimidine-, oxaliplatin- and irinotecan-containing chemotherapy; its letter states that the approval fulfils the Subpart E (accelerated approval) commitments of the 2006 licence, and names the updated results of the PRIME trial among the studies used to verify clinical benefit.

Why it mattered

The PRIME reports tied the effect of an anti-EGFR antibody to tumour genotype: KRAS exon 2 mutations, then other RAS mutations, were associated with no benefit or worse outcomes with panitumumab-FOLFOX4, and the US indication moved from KRAS exon 2 to RAS wild-type. FIRE-3 posed the head-to-head question between cetuximab and bevacizumab with FOLFIRI, which its report describes as unknown.

Caveats

  • The PRIME RAS analysis was prospective-retrospective and the overall rate of ascertainment of RAS status was 90%; the 512-patient RAS wild-type subgroup is not the trial's primary comparison.
  • In PRIME's KRAS wild-type stratum, overall survival did not differ significantly in the 2010 report (HR 0.83, 95% CI 0.67-1.02, P=0.072).
  • FIRE-3's primary endpoint (objective response) did not differ significantly between cetuximab and bevacizumab; the longer overall survival with cetuximab is a non-primary finding of an open-label trial, and the later head-to-head CALGB/SWOG trial did not show a significant overall survival difference (see the primary tumour side milestone).
  • COIN, which added cetuximab to oxaliplatin-based chemotherapy, found no overall survival or progression-free survival benefit in KRAS wild-type tumours (median OS 17.9 months with chemotherapy alone vs 17.0 months with cetuximab, HR 1.04); first-line anti-EGFR results differ across trials and chemotherapy backbones.
  • OPUS is a randomised phase 2 study whose primary endpoint was response rate; its RAS wild-type subgroup (87 patients) was analysed post hoc, is small, and its report states that small numbers limited precise estimates.
  • Each trial's figures are its own: the chemotherapy backbones (FOLFOX4, FOLFIRI), the depth of RAS testing and the populations differ, so the results are not directly comparable.