Treating itColorectal cancer

Colonoscopy, FIT, Cologuard, Shield blood test from age 45.

Surgery; adjuvant chemotherapy guided by risk and (in trials) ctDNA; exercise programme.

Doublet/triplet chemotherapy + biologic by genotype; targeted combinations for BRAF, KRAS G12C, HER2.

Checkpoint inhibitors; organ preservation in rectal cancer.

Colonoscopy every 10 years, annual FIT, multitarget stool DNA every 3 years, CT colonography, or the Shield blood test every 3 years; start at 45 (USPSTF 2021). Lynch carriers: colonoscopy every 1-2 years from age 20-25.

Surgical resection; observation for stage I and low-risk stage II. High-risk stage II (T4, obstruction, <12 nodes, LVI): consider 3-6 months fluoropyrimidine ± oxaliplatin; ctDNA-negative patients can safely omit (DYNAMIC). dMMR stage II derives no benefit from 5-FU alone.

Resection then adjuvant CAPOX for 3 months (T1-3 N1, low risk) or FOLFOX/CAPOX for 6 months (T4 or N2), per IDEA.

Resection then FOLFOX + atezolizumab for 12 months (ATOMIC, 3-year DFS 86% vs 77%). Neoadjuvant nivolumab + ipilimumab (NICHE-2) is an alternative under evaluation.

Total neoadjuvant therapy: short-course radiation or long-course chemoradiation plus FOLFOX/CAPOX, then TME surgery or watch-and-wait for clinical complete response (OPRA). Non-operative management for select cCR.

Dostarlimab monotherapy for 6 months with organ preservation (MSK cohort 100% cCR; AZUR-1 registrational, PDUFA February 2027). Chemoradiation and surgery reserved for non-responders.

Pembrolizumab (KEYNOTE-177) or nivolumab + ipilimumab (CheckMate 8HW, PFS 54 months); nivolumab monotherapy alternative. Chemotherapy reserved for IO-refractory disease.

FOLFOX or FOLFIRI + panitumumab or cetuximab (PARADIGM OS 37.9 months); bevacizumab if anti-EGFR contraindicated. Maintenance fluoropyrimidine ± biologic after induction.

FOLFOX, FOLFIRI, or FOLFOXIRI + bevacizumab; anti-EGFR contraindicated. KRAS G12D and other alleles: RAS(ON) inhibitors in trials.

First line: encorafenib + cetuximab + mFOLFOX6 or FOLFIRI (BREAKWATER, OS 30.3 vs 15.1 months). Later line: encorafenib + cetuximab (BEACON).

Sotorasib + panitumumab (CodeBreaK 300) or adagrasib + cetuximab after chemotherapy; first-line trials ongoing.

Tucatinib + trastuzumab (MOUNTAINEER) or trastuzumab deruxtecan 5.4 mg/kg (DESTINY-CRC02) after chemotherapy; MOUNTAINEER-03 tests first-line use.

Trifluridine/tipiracil + bevacizumab (SUNLIGHT, OS 10.8 months), then fruquintinib (FRESCO-2) or regorafenib; NTRK inhibitor if fusion; clinical trials.

Resection, thermal ablation, or SBRT with curative intent after multidisciplinary review; perioperative chemotherapy; 5-year survival 30-50% after complete resection of liver metastases.

Removing the cancer with a rim of healthy bowel, and joining the two ends back together, is what cures most bowel cancer. The operation depends on where the tumour sits. Macmillan names the colon operations as a right or left hemi-colectomy, a sigmoid colectomy (also called a high anterior resection), a transverse colectomy or a total colectomy, and the rectal ones as a local excision for a very small stage 1 cancer, an anterior resection or low anterior resection for the upper or middle rectum, usually with a total mesorectal excision, and it says that if the cancer is very low in the rectum and close to the anus you are more likely to need a permanent stoma. NICE NG151 recommends laparoscopic (keyhole) resection as an alternative to open resection for both colon and rectal cancer when both are suitable, says to consider open surgery for locally advanced tumours or after previous abdominal or pelvic surgery, and to use robotic surgery only within established programmes with audited outcomes. It sets volume floors for rectal surgery: at least 10 major resections a year per hospital and at least 5 per surgeon. On the stoma, NICE asks teams to advise people of the likelihood of having one, why it might be necessary and for how long it might be needed, and to have a trained stoma professional provide information on care and on learning to live with it. Bowel Cancer UK puts it plainly: a reversible stoma is formed to let the join heal and a permanent one when the two ends cannot be joined, and the surgeon may not know for certain until the operation has started. A colostomy is formed from the large bowel and the output is more solid; an ileostomy is formed from the small bowel, the output is looser and more fluid and salt are lost. The NHS says recovery from a colostomy usually takes about 8 weeks, with no heavy lifting in that time, and that permanent colostomy supplies come free on prescription. Prehabilitation before the operation and an enhanced recovery programme after it are both worth asking about: NICE says to explain what recovery protocols involve and their value, and Bowel Cancer UK says prehabilitation reduces anxiety, improves fitness and can shorten the stay.

Chemotherapy after the operation is given to kill cells that have already left the bowel but are too few to see, and the whole decision is about how large that risk is against what the treatment costs you. For stage 3 (cancer in the lymph nodes), NICE NG151 is explicit and the same for colon and for rectal cancer treated with short-course radiotherapy or no preoperative treatment: offer capecitabine with oxaliplatin (CAPOX) for 3 months, or if that is not suitable either oxaliplatin with fluorouracil and folinic acid (FOLFOX) for 3 to 6 months, or a single-agent fluoropyrimidine such as capecitabine for 6 months. It then says to base the choice on the person's histopathology (giving pT1 to T3 with pN1 against pT4 or pN2 as the example), performance status, personal preferences, comorbidities and age, which is the guideline's way of saying that three months of a doublet and six months of a single tablet are both defensible and the balance is yours to weigh. For stage 2 (no nodes involved) NICE NG151 makes no adjuvant recommendation at all, so the conversation runs on individual risk features, on the mismatch repair result, and increasingly on whether circulating tumour DNA is detectable after surgery; the ctDNA-guided studies on this record (DYNAMIC and the others) are where that question is being settled, and a trial is a reasonable answer. Two practical points NICE puts in writing: monitor prolonged sensory symptoms after platinum chemotherapy, because they are a sign the dose needs reducing to prevent permanent neuropathy, and a DPD blood test comes before fluorouracil or capecitabine because low DPD raises the risk of severe toxicity. The OnCo decision aid at /tools/colorectal-adjuvant-chemotherapy/ lays the NICE wording out by site, stage and fitness for oxaliplatin.

Rectal cancer is the part of bowel cancer where the order of treatment is a real choice. NICE NG151 says not to offer preoperative radiotherapy for early rectal cancer (cT1 to T2, cN0, M0) outside a clinical trial, and to offer preoperative radiotherapy or chemoradiotherapy where the cancer is cT1 to T2 with involved nodes, or cT3 to T4 with any node status. Two forms are used: a short course of radiotherapy on its own, and a longer course of radiotherapy given together with chemotherapy (Macmillan says the drug most commonly used is capecitabine tablets, started on the first morning of radiotherapy and taken throughout it, with a DPD test first), and in recent practice both are often followed by several months of chemotherapy before surgery (total neoadjuvant therapy, tested in RAPIDO, PRODIGE 23 and OPRA on this record; PROSPECT asked the opposite question, whether chemotherapy alone can replace radiotherapy in selected tumours). The reason this matters to a patient is that treatment before surgery shrinks the tumour, improves the chance of a clear margin, and in a minority makes the tumour disappear altogether on examination, MRI and endoscopy. NICE NG151 covers that case directly: inform people with a complete clinical and radiological response who wish to defer surgery that there is a risk of recurrence, and that there are no prognostic factors to guide selection for deferral; for those who choose to defer, encourage participation in a clinical trial and ensure data is collected via a national registry. Watch and wait is therefore not less treatment but a different commitment, to frequent examination and scans for years. For early rectal cancer NICE offers a shared decision between transanal excision, endoscopic submucosal dissection and total mesorectal excision, and gives a table comparing them on exactly the points a patient asks about: whether bowel is removed, whether a stoma is possible, hospital stay, scarring and the complications of each.

When bowel cancer has spread, the first treatment is chosen by three test results and one piece of anatomy. NICE NG151 says to test for RAS and BRAF V600E mutations in all people with metastatic colorectal cancer suitable for systemic anticancer treatment, and separately recommends immunotherapy where the tumour is MSI-high or mismatch repair deficient. If RAS is wild-type, NICE recommends cetuximab or panitumumab with FOLFOX or FOLFIRI; if RAS is mutated, those antibodies do not work and bevacizumab with fluoropyrimidine-based chemotherapy is recommended where targeted treatment or immunotherapy is unsuitable. The anatomy is sidedness: cancers that start on the left (descending colon, sigmoid, rectum) respond better to the EGFR antibodies than cancers that start on the right, which is why the record's own rows put anti-EGFR treatment with left-sided RAS wild-type disease and bevacizumab with right-sided or RAS-mutant disease (PARADIGM tested that question head to head; CRYSTAL and FIRE-3 are the older comparisons). BRAF V600E is its own group, with encorafenib and cetuximab added to chemotherapy (BREAKWATER). What this means at the appointment is that the first question is not which drug but which result, and that a treatment plan made before the molecular report is back may change. Two further decisions sit alongside: whether the primary tumour should be removed when it is causing no symptoms (NICE says consider it, and supplies a table of advantages and disadvantages to share), and when to take a break from oxaliplatin before the nerve damage becomes permanent.

Bowel cancer that has spread to the liver is one of the few settings in which surgery for metastatic disease is done with the intention of cure, and a meaningful number of people live many years after it. NICE NG151 says to consider resection, either simultaneous with the bowel operation or sequential, after discussion by a multidisciplinary team with expertise in resection of disease in all involved sites, and to consider perioperative systemic anticancer therapy where liver resection is a suitable treatment. Where the liver disease is not resectable, it says to consider chemotherapy with local ablative techniques after discussion by a specialist multidisciplinary team, and not to offer selective internal radiation therapy as first-line treatment for liver metastases unsuitable for local treatment except under the arrangements its HealthTech guidance sets out. The same logic extends to the lungs: consider metastasectomy, ablation or stereotactic body radiation therapy for lung metastases suitable for local treatment, after discussion by a team including a thoracic surgeon and a specialist in non-surgical ablation, and consider a biopsy for a single lung lesion to exclude a primary lung cancer. For disease limited to the peritoneum, NICE says to offer systemic therapy and, within the multidisciplinary team, to discuss referral to a nationally commissioned specialist centre to consider cytoreductive surgery with heated intraperitoneal chemotherapy. The practical consequence for a patient is that the answer to whether an operation is possible depends on which team is looking, so asking whether a specialist liver or peritoneal unit has seen the scans is a fair question, and a second opinion is a recognised step (Bowel Cancer UK has a page on it).

A small minority of bowel cancers that have spread have lost the mismatch repair machinery that proof-reads DNA, which leaves the tumour carrying very many mutations and makes it visible to the immune system. For those people, immunotherapy has replaced chemotherapy as the first treatment. NICE NG151 recommends nivolumab with ipilimumab as an option for untreated unresectable or metastatic colorectal cancer with high microsatellite instability or mismatch repair deficiency, and pembrolizumab as an option for untreated metastatic disease with the same markers, stopped after 2 years or earlier if the disease progresses; after fluoropyrimidine-based combination chemotherapy it recommends nivolumab with ipilimumab, and pembrolizumab only if that combination cannot be used. KEYNOTE-177 and CheckMate 8HW are the trials behind those recommendations and are on this record. The same biology is being tested before surgery, where a few weeks of immunotherapy has cleared the tumour in a large fraction of mismatch repair deficient colon cancers (NICHE-2) and of rectal cancers (the dostarlimab work and the registrational AZUR-1), but outside those trials NICE does not recommend it, so the question at the clinic is whether a trial is open. Everything therefore turns on the test: ask whether mismatch repair immunohistochemistry or microsatellite instability testing has been done and what the result was, because a mismatch repair deficient result also starts the Lynch syndrome pathway for your family.

When the cancer grows through the first and second lines, there is still a recognised sequence, and knowing it in advance makes each step less sudden. NICE NG151 points to its appraisals for metastatic colorectal cancer previously treated with fluoropyrimidine-based chemotherapy, anti-VEGF or anti-EGFR therapy: trifluridine with tipiracil plus bevacizumab after 2 systemic treatments (the SUNLIGHT combination), fruquintinib after 2 systemic treatments if trifluridine with tipiracil plus bevacizumab is not suitable (FRESCO-2), regorafenib, and trifluridine with tipiracil alone. It also lists what is not recommended at this point, including aflibercept with irinotecan and fluorouracil after oxaliplatin, and cetuximab or panitumumab monotherapy after first-line chemotherapy. Where a rarer change is present the sequence changes: encorafenib with cetuximab for BRAF V600E after previous systemic treatment, larotrectinib through the Cancer Drugs Fund for NTRK fusion-positive tumours when there are no other satisfactory options, and the HER2 and KRAS G12C options on this record's other rows. These drugs are given to hold the cancer and protect quality of life rather than to cure, so the honest questions are what each is likely to buy, what it costs in side effects and hospital visits, how progression will be recognised, and what the stopping rule is. Bowel Cancer UK has pages on treating advanced bowel cancer and on taking a break from treatment; both are legitimate choices, and so is a trial.

Around 3 in 100 UK bowel cancers are caused by Lynch syndrome, an inherited fault in one of the mismatch repair genes, and Bowel Cancer UK says testing should be offered to everyone at diagnosis because the result changes treatment as well as family risk. The route runs in order: the tumour is stained for the mismatch repair proteins (or tested for microsatellite instability); where MLH1 is lost, the laboratory checks for MLH1 promoter methylation, which usually means the loss is sporadic rather than inherited; and where the pattern still suggests an inherited cause, a germline blood test is offered through the genomics service. A positive result has three consequences. For you, immunotherapy becomes an option in advanced disease, and lifelong colonoscopy surveillance begins. For your relatives, each child, brother and sister has a one in two chance of carrying the same change, and Bowel Cancer UK says they should be offered the same test, which it calls cascade testing; it estimates that most of the 175,000 to 200,000 people in the UK with Lynch syndrome do not know. For prevention, NICE NG151 says to consider aspirin, taken daily and for a period of more than 2 years, to reduce the risk of colorectal cancer in people with Lynch syndrome, noting that in January 2020 this was an off-label use and that NICE has produced a patient decision aid to support the discussion. Bowel Cancer UK says surveillance means colonoscopy every 18 months to two years and can reduce the chance of dying from bowel cancer by as much as 72 percent.

At several points in bowel cancer a trial is a reasonable choice beside standard treatment rather than a last resort, and NICE NG151 says so itself in two places: preoperative radiotherapy for early rectal cancer should only happen within a clinical trial, and people who defer surgery after a complete response should be encouraged to take part in a trial with data collected via a national registry. The open questions this record follows are the ones a trial would answer for you: whether circulating tumour DNA after surgery should decide who has chemotherapy and for how long; whether immunotherapy before surgery can replace an operation in mismatch repair deficient colon and rectal cancer; whether the newer RAS inhibitors help the large majority whose tumours are RAS-mutant; and how to make immunotherapy work in microsatellite-stable disease, where it so far does not. Bowel Cancer UK says some trials test new treatments and others test new ways of using existing treatments, and that if you would like to take part you should ask your healthcare team whether you are suitable for any current trials in your area; the NHS says you can ask your doctor or a patient organisation about trials you may be eligible to join, and that you can choose to leave at any point without giving a reason and without it affecting the care you receive. Practical points worth settling before you agree: how many extra visits, scans or biopsies are involved, whether travel is reimbursed, whether the trial is open at another hospital if not here, and what the standard alternative would be if you decline.

Palliative care is symptom control and support, and it is not a stage of the illness. Bowel Cancer UK says palliative care may start when you are diagnosed with advanced bowel cancer, or after treatment has stopped working, and that its aim is to improve quality of life; end of life care is a later part of the same service, usually in the last year of life. The NHS says that where bowel cancer cannot be cured you will be referred to a symptom control or palliative care team, who work with you to manage symptoms and help you and your loved ones get other support. A randomised trial in another cancer (Temel 2010) found that early palliative care alongside cancer treatment improved quality of life and mood, which is why the referral belongs alongside treatment rather than after it. In bowel cancer the concrete reasons are specific: pain from pelvic or liver disease, a bowel that is becoming obstructed, high stoma output and dehydration, poor appetite and weight loss, fatigue, low mood, sleep, and the practical questions about work, money and care at home. Bowel Cancer UK says you can choose where you would prefer to be cared for and where you wish to die, that care can be at home with the GP and community nurses, in a hospice (for a day or for a stay), in hospital or in a care home, and that your choices can be documented by your healthcare team and kept with your medical records so that the people close to you know them. Marie Curie sets out what palliative care covers and who provides it, and its support line and nurses are available to families as well as patients.

Resection with complete mesocolic excision and at least twelve lymph nodes examined; observation afterwards. Adjuvant chemotherapy is not given for stage I. In stage II the absolute gain from fluorouracil and folinic acid is small: QUASAR (3,239 patients, 2,963 with stage II disease) found a relative risk of death of 0.82 and of recurrence 0.78, translating to about 3.6 percent absolute survival. DYNAMIC showed that a negative circulating tumour DNA result four to seven weeks after surgery identifies patients who can safely have no chemotherapy, cutting adjuvant use from 27.9 to 15.3 percent without worse recurrence-free survival. Mismatch-repair deficient stage II disease gains nothing from fluorouracil alone and is observed.

Resection then a discussion about three to six months of a fluoropyrimidine with or without oxaliplatin, weighing T4 disease, obstruction or perforation, fewer than twelve nodes examined, lymphovascular or perineural invasion and poor differentiation against the neuropathy risk. SCOT included high-risk stage II patients and found three months of oxaliplatin-containing chemotherapy non-inferior to six (three-year disease-free survival 76.7 against 77.1 percent) with grade 2 or worse neuropathy halved. Mismatch-repair deficiency remains a reason not to give fluorouracil alone.

Resection then adjuvant oxaliplatin-fluoropyrimidine chemotherapy. MOSAIC established it (three-year disease-free survival 78.2 against 72.9 percent with fluorouracil alone, hazard ratio 0.77; six-year overall survival gain confined to stage III). Duration follows IDEA: three months of CAPOX for T1-3 N1 disease (three-year disease-free survival 83.1 against 83.3 percent for six months) and six months of FOLFOX or CAPOX for T4 or N2. Cetuximab must not be added: N0147 and PETACC-8 both found no benefit and more toxicity.

Resection then FOLFOX with atezolizumab for twelve months (ATOMIC, three-year disease-free survival 86 against 77 percent), the first adjuvant immunotherapy success in colorectal cancer. Neoadjuvant nivolumab with ipilimumab is the alternative under test: NICHE-2 produced a 68 percent pathological complete response rate and AZUR-2 (NCT05855200, 892 patients, primary completion March 2029) is the phase 3 that would make perioperative dostarlimab standard.

Six weeks of preoperative oxaliplatin-fluoropyrimidine chemotherapy followed by surgery and eighteen weeks more is an option for radiologically staged T3 with deep invasion or T4 disease. FOxTROT (1,053 patients) found residual or recurrent disease within two years in 16.9 against 21.5 percent with surgery first (rate ratio 0.72), with fewer serious postoperative complications and 4.3 percent needing expedited surgery for obstruction. Panitumumab added nothing and mismatch-repair deficient tumours gained little, so they are better served by neoadjuvant immunotherapy in a trial.

Total mesorectal excision is the operation (Dutch TME trial) and preoperative rather than postoperative radiotherapy is the sequence (CAO/ARO/AIO-94: five-year local relapse 6 against 13 percent, no survival difference). High-risk disease (cT4, extramural vascular invasion, cN2, threatened mesorectal fascia, lateral nodes) is treated with total neoadjuvant therapy: short-course 5 x 5 Gy then CAPOX or FOLFOX before surgery (RAPIDO: three-year disease-related treatment failure 23.7 against 30.4 percent, hazard ratio 0.75; STELLAR: three-year disease-free survival 64.5 against 62.3 percent) or long-course chemoradiotherapy with FOLFIRINOX before it (PRODIGE 23: three-year disease-free survival 76 against 69 percent, hazard ratio 0.69). Delaying surgery four to eight weeks after short-course radiotherapy is safe (Stockholm III). For sphincter-sparing candidates who do not need radiotherapy, PROSPECT showed neoadjuvant FOLFOX with selective chemoradiotherapy is non-inferior (five-year disease-free survival 80.8 against 78.6 percent).

Consolidation chemotherapy after chemoradiotherapy, not induction before it, is the sequence that preserves the rectum: OPRA gave five-year TME-free survival of 54 against 39 percent, and CAO/ARO/AIO-12 found pathological complete response 25 against 17 percent with the same ordering. Patients with a clinical complete response on digital examination, endoscopy and MRI are offered watch and wait with endoscopy and MRI every three to four months. The International Watch & Wait Database (880 patients with a clinical complete response from 47 institutes) records two-year local regrowth of 25.2 percent, 88 percent of it within two years and 97 percent in the bowel wall where salvage surgery is possible, with five-year overall survival 85 percent and disease-specific survival 94 percent. In mismatch-repair deficient rectal cancer, dostarlimab for six months has produced complete clinical responses without surgery and AZUR-1 is the registrational trial.

Immunotherapy, not chemotherapy: pembrolizumab (KEYNOTE-177) or nivolumab with ipilimumab (CheckMate 8HW). In England NICE TA709 recommends pembrolizumab stopped at two years and TA1065 recommends nivolumab with ipilimumab; TA914 places pembrolizumab after fluoropyrimidine therapy only where nivolumab with ipilimumab is unsuitable. The FDA approved nivolumab with ipilimumab for this indication on 8 April 2025. Universal mismatch-repair or MSI testing at diagnosis is what makes this route possible.

FOLFOX or FOLFIRI with panitumumab or cetuximab. PARADIGM showed panitumumab beats bevacizumab in left-sided RAS wild-type disease; FIRE-3 found overall survival 28.7 against 25.0 months for cetuximab against bevacizumab with FOLFIRI, while CALGB/SWOG 80405 found no overall difference (30.0 against 29.0 months) before sidedness was taken into account. Extended RAS testing of KRAS and NRAS exons 2, 3 and 4 is mandatory: PRIME showed patients with non-exon-2 RAS mutations do worse with panitumumab. In England TA439 funds cetuximab or panitumumab with FOLFOX or FOLFIRI first line only.

FOLFOX, FOLFIRI or FOLFOXIRI with bevacizumab; EGFR antibodies are contraindicated. FOLFOXIRI plus bevacizumab gives a higher response rate and longer progression-free survival than a doublet (TRIBE: 12.1 against 9.7 months; TRIBE2 progression-free survival 2 of 19.2 against 16.4 months) at the cost of grade 3-4 neutropenia in about half, and is preferred where deep response matters. After six cycles, maintenance with a fluoropyrimidine and bevacizumab beats observation (CAIRO3: progression-free survival 2 of 11.7 against 8.5 months). In England bevacizumab is funded under TA1136 only when targeted treatment or immunotherapy is unsuitable.

First line encorafenib with cetuximab and mFOLFOX6 (BREAKWATER), which the FDA approved on 20 December 2024 and converted to traditional approval on 24 February 2026. After previous treatment, encorafenib with cetuximab (BEACON CRC), which NICE recommends in England under TA668. The mechanism was established by SWOG S1406: BRAF blockade alone fails because it releases feedback activation of EGFR, so the EGFR antibody is not optional. There is no NICE recommendation for the first-line triplet.

Sotorasib with panitumumab (CodeBreaK 300, FDA approval 16 January 2025) or adagrasib with cetuximab (KRYSTAL-1 colorectal cohort, accelerated approval 21 June 2024) after fluoropyrimidine, oxaliplatin and irinotecan. The confirmatory trials are KRYSTAL-10 (NCT04793958, 461 patients, active but closed) and, for first-line use, CodeBreaK 301 (NCT06252649, 450 patients, recruiting). Neither combination has a NICE recommendation, so they are not routinely funded in England.

Tucatinib with trastuzumab (MOUNTAINEER, confirmed objective response 38.1 percent; FDA accelerated approval 19 January 2023) or trastuzumab deruxtecan at 5.4 mg/kg (DESTINY-CRC02, response 37.8 percent; DESTINY-CRC01 gave 45.3 percent at the higher 6.4 mg/kg dose with interstitial lung disease in 6 percent and two deaths). The tumour-agnostic HER2 IHC 3+ accelerated approval of trastuzumab deruxtecan came on 5 April 2024. MOUNTAINEER-03 (NCT05253651, 400 patients, primary completion December 2027) is testing first-line use. No NICE recommendation exists for either regimen in colorectal cancer.

NTRK fusions occur in well under 1 percent of colorectal cancers, almost always in right-sided, mismatch-repair deficient, RAS and BRAF wild-type tumours, and are treated with larotrectinib (NICE TA630) or entrectinib (NICE TA644), both recommended tumour-agnostically in England. RET fusions are rarer still and are treated with selpercatinib under its tumour-agnostic approval. POLE or POLD1 exonuclease-domain mutations produce an ultramutated, immunotherapy-responsive tumour that is microsatellite-stable, so it is missed unless sequencing is done. All three are reasons to send comprehensive sequencing rather than a small hotspot panel in metastatic disease.

Trifluridine-tipiracil with bevacizumab (SUNLIGHT: overall survival 10.8 against 7.5 months; FDA approval 2 August 2023; NICE TA1008), then fruquintinib (FRESCO-2: 7.4 against 4.8 months; FDA approval 8 November 2023; NICE TA1079 only where trifluridine-tipiracil with bevacizumab is unsuitable) or regorafenib (CORRECT: 6.4 against 5.0 months; NICE TA866). Second-line anti-angiogenic options after an oxaliplatin regimen are ziv-aflibercept (VELOUR: 13.50 against 12.06 months, not recommended by NICE in TA307) and ramucirumab (RAISE: 13.3 against 11.7 months, no NICE recommendation). Anti-EGFR rechallenge guided by circulating tumour DNA is promising and unproven in phase 3.

Resistance to EGFR antibodies is universal and arises through RAS, BRAF and EGFR ectodomain mutant clones that expand under treatment and decay once it stops, which is why rechallenge is worth trying and why it must be selected by a blood test rather than by time alone. CHRONOS screened 52 patients with circulating tumour DNA, excluded the 31 percent with a resistance mutation and rechallenged the rest with panitumumab alone: 30 percent responded and 63 percent had disease control. CAVE rechallenged with cetuximab and avelumab and found median overall survival 17.3 months where baseline plasma RAS and BRAF were wild-type against 10.4 months where they were mutated (hazard ratio 0.49). Both trials are single-arm, so rechallenge is an option to consider in a trial or after discussion, not a standard.

Ninety-five percent of metastatic colorectal cancers are microsatellite-stable and do not respond to PD-1 blockade; IMblaze370 confirmed that adding a MEK inhibitor does not change this (overall survival 8.87 against 8.51 months with regorafenib, hazard ratio 1.00). The most promising current approach is Fc-enhanced CTLA-4 blockade restricted to patients without liver metastases: botensilimab with balstilimab gave an objective response rate of 21 percent, median overall survival 21.2 months and 36-month overall survival 33 percent in 123 previously treated patients, uncontrolled. Liver metastases deplete tumour-specific CD8 T cells, which is the reason for the restriction and the reason results in unselected populations have been so poor.

Resection with curative intent where all disease can be removed with an adequate remnant; perioperative FOLFOX is the reference, although EORTC 40983 found the progression-free survival gain did not become a survival gain at 8.5 years (61.3 against 54.3 months, hazard ratio 0.88). Thermal ablation is the only local treatment with randomised survival evidence: CLOCC gave eight-year overall survival of 35.9 against 8.9 percent when radiofrequency ablation was added to systemic treatment. For initially unresectable liver metastases, CAIRO5 chose FOLFOXIRI with bevacizumab for right-sided or RAS/BRAF-mutant tumours and found panitumumab no better than bevacizumab for left-sided wild-type tumours, with resectability reviewed by a surgical panel every two months. Cetuximab must not be added around liver resection: New EPOC found median overall survival 55.4 against 81.0 months with it.

Complete cytoreductive surgery in selected patients with a Peritoneal Cancer Index of 25 or less gives a median overall survival above three years, but adding heated intraperitoneal oxaliplatin does not: PRODIGE 7 found 41.7 against 41.2 months with more late grade 3 or worse complications. HIPEC also failed as prophylaxis in T4 or perforated colon cancer (COLOPEC) and as systematic second-look surgery (PROPHYLOCHIP). Selection for surgery, and systemic chemotherapy, are what change the outcome; other agents and delivery methods, including pressurised intraperitoneal aerosol chemotherapy, remain investigational.

A tumour-informed assay four to seven weeks after resection is the strongest single predictor of recurrence: GALAXY within CIRCULATE-Japan and DYNAMIC-III both show a large separation (DYNAMIC-III three-year recurrence-free survival 87 percent if negative against 49 percent if positive). De-escalation on a negative result is proven only in stage II (DYNAMIC, adjuvant chemotherapy use 15.3 against 27.9 percent with equal recurrence-free survival); in stage III, DYNAMIC-III could not show non-inferiority. Escalation on a positive result has failed twice: ALTAIR missed its disease-free survival endpoint (9.30 against 5.55 months, hazard ratio 0.79, p=0.107) and DYNAMIC-III's escalation arm gave no benefit. CIRCULATE-US (NRG-GI008) and TRACC are the trials still running.

Adjuvant cetuximab (N0147, PETACC-8) gave no benefit and more toxicity; cetuximab around liver resection shortened life (New EPOC, 55.4 against 81.0 months). Oxaliplatin HIPEC failed three times (PRODIGE 7, COLOPEC, PROPHYLOCHIP). Checkpoint inhibition with a MEK inhibitor failed in microsatellite-stable disease (IMblaze370, hazard ratio 1.00 against regorafenib). Treating molecular recurrence with late-line chemotherapy failed (ALTAIR). In access, NICE did not recommend ziv-aflibercept (TA307) despite a positive phase 3 trial.

Funded in England: cetuximab and panitumumab first line in RAS wild-type disease (TA439); pembrolizumab first line for MSI-high or dMMR disease (TA709) and after chemotherapy where nivolumab with ipilimumab is unsuitable (TA914); nivolumab with ipilimumab first line (TA1065) and after previous treatment (TA716); encorafenib with cetuximab after previous treatment in BRAF V600E disease (TA668); trifluridine-tipiracil (TA405) and with bevacizumab after two lines (TA1008); regorafenib (TA866); fruquintinib at third line or later where trifluridine-tipiracil with bevacizumab is unsuitable (TA1079); bevacizumab with fluoropyrimidine chemotherapy first and second line where targeted treatment or immunotherapy is unsuitable (TA1136); larotrectinib (TA630) and entrectinib (TA644) for NTRK fusions. Not funded: ziv-aflibercept (TA307, not recommended), and, for want of any appraisal, tucatinib with trastuzumab, trastuzumab deruxtecan in colorectal cancer, sotorasib with panitumumab, adagrasib with cetuximab and first-line encorafenib with chemotherapy. The clinical guideline is NG151.

A self-expanding metal stent at colonoscopy relieves a left-sided obstruction within a day or two, either as a bridge to elective resection without a stoma or as definitive palliation; perforation, migration and re-obstruction are the risks and stenting is avoided during bevacizumab treatment. A defunctioning or end colostomy is preferred for right-sided or multi-level obstruction, peritoneal disease and a longer prognosis; right-sided obstruction is usually resected with primary anastomosis. Inoperable malignant bowel obstruction is managed medically with a syringe driver: an anti-emetic, hyoscine butylbromide or octreotide, dexamethasone and opioid analgesia, with a venting gastrostomy if vomiting persists.

Liver capsule pain from bulky metastases is a position-dependent right upper quadrant pain that usually responds to dexamethasone and, when it does not, to a short course of palliative radiotherapy to the liver. Bleeding, discharge and pain from an unresectable rectal tumour respond to hypofractionated palliative radiotherapy; tenesmus is the hardest symptom and may need a nerve block or a stoma. Opioids are titrated with laxatives, and anaemia from chronic blood loss is treated with iron or transfusion. Early integrated palliative care alongside oncology from the diagnosis of incurable disease is recommended, and dietetic review matters where a stoma or short bowel follows surgery.

Quantitative faecal immunochemical testing in primary care to decide referral; colonoscopy with biopsy, or computed tomography colonography where colonoscopy is not possible; computed tomography of chest, abdomen and pelvis for everyone and high-resolution pelvic magnetic resonance imaging for every rectal cancer; RAS and BRAF V600E testing in metastatic disease before systemic therapy; mismatch repair testing at diagnosis, which also finds Lynch syndrome; baseline carcinoembryonic antigen.

Stenting for people being treated with palliative intent; either stenting or emergency surgery when potentially curative treatment is suitable. In CReST, stenting as a bridge to elective surgery relieved obstruction in 82.4 percent and cut stoma formation from 67.9 to 47.5 percent in patients treated with curative intent, with no difference in 30-day mortality, hospital stay, 3-year recurrence or survival.

Remove adenomas at colonoscopy and put the person on a surveillance interval set by what was found; consider daily aspirin for more than two years in Lynch syndrome. Population screening from 50 in England and 45 in the United States. Diet, weight, alcohol, smoking and physical activity account for the 54 percent of UK cases Cancer Research UK judges preventable.

Follow up for detection of local recurrence and distant metastases for the first three years, including serum carcinoembryonic antigen and computed tomography of the chest, abdomen and pelvis; a clearance colonoscopy at one year and a surveillance colonoscopy three years after that.

Offer a quantitative faecal immunochemical test to the symptom groups NICE lists and refer on the suspected cancer pathway at 10 micrograms of haemoglobin per gram of faeces or above; refer a rectal mass without testing first; safety net people who do not return a sample or test below the threshold, and do not let a low result delay referral where clinical concern is strong.