DataColorectal cancer

Direct links plus the targets, companies, and technologies of this cancer's products.

• Living with bowel cancer, care and decisions: This is orientation from public patient pages and guidelines, not advice for your case: your own team's instructions and 24-hour number come first. The ten decision rows above, the question sets for the surgeon, oncologist, stoma nurse, genomics and palliative care appointments, the first 60 days checklist and the red cards were written from NICE NG151 and the NHS, Bowel Cancer UK, Macmillan, Cancer Research UK, Colostomy UK, Maggie's and Marie Curie patient pages, all read on 24 September 2026.
• The science in detail, 1 of 7, in full; the one-sentence version is on the overview. Genomic landscape. WNT signalling is altered in 93% of tumours in the TCGA analysis and 96% once intronic APC splice events and large in-frame CTNNB1 deletions are counted: APC mutated in 58 to 77% by cohort, with RNF43 5 to 12%, CTNNB1 5 to 7%, AMER1 6 to 13%, TCF7L2 7 to 15% and RSPO2 or RSPO3 fusions covering most of the APC-wild-type remainder. TP53 52 to 73%, KRAS 40 to 44%, PIK3CA 20 to 28%, FBXW7 13 to 17%, SMAD4 12 to 16%, ARID1A 9 to 13%, SOX9 10 to 12%, NRAS 4 to 9%, BRAF V600E 6 to 18%, B2M 3 to 7%, ERBB2 amplification 2 to 3% and mutation 4 to 6%, MYC amplification 4 to 5%, IGF2 amplification about 2%, and fusions of NTRK, RET and ALK each at or below 0.3%. Every figure and its cohort is in the molecular table; the cBioPortal rows were computed on the TCGA PanCancer Atlas and 2012 deposits, the DFCI prospective exome cohort, the Genentech discovery set, two MSK-IMPACT cohorts (1,134 and 7,237 samples), the MSK early-onset comparison and the Chinese ChangKang project.
• The science in detail, 2 of 7, in full; the one-sentence version is on the overview. Two instability classes. About 16% of colorectal cancers are hypermutated: three-quarters through mismatch repair failure, usually sporadic MLH1 promoter silencing, and one-quarter through somatic mismatch repair gene or POLE mutation. The other 84% are chromosomally unstable, with a persistent segregation defect running above 10^-2 chromosome gains or losses per division and the characteristic 18q, 17p and 8p losses and 8q, 13q and 20q gains. The two classes are near-exclusive, which is why SMAD4 and TP53 loss are rare in MSI-high tumours and why coding-microsatellite frameshift targets (TGFBR2 38.6% against 1.8%, RNF43 63.8% against 3.7%, ARID1A 65.2% against 6.3%, B2M 31.0% against 1.5%) are almost confined to them. Median tumour mutational burden separates three regimes: 5.7 per megabase microsatellite stable, 56.9 MSI-high and 172.1 in POLE exonuclease hotspot tumours.
• The science in detail, 3 of 7, in full; the one-sentence version is on the overview. Sidedness as biology. In the 3,483 MSK-IMPACT samples with a named subsite, microsatellite instability was 24.3% on the right against 5.2% on the left, BRAF V600E 19.2% against 2.9%, KRAS 51.9% against 39.7%, PIK3CA 32.4% against 15.5%, APC 65.0% against 77.6% and TP53 56.7% against 79.2%. Right-sided microsatellite-stable tumours carry mitogenic pathway mutations; left-sided ones often carry none and depend on ligand instead, which is the mechanism behind the treatment rule. Across six randomised trials in 2,159 RAS wild-type patients, adding an EGFR antibody improved overall survival on the left (hazard ratio 0.75) and not on the right (1.12), interaction p less than 0.001, while right-sided disease was worse in every arm (overall survival hazard ratio 2.03 in the control arms). The caution is that the bowel is a gradient: RAS mutation falls from 70% at the caecum to 43% at the hepatic flexure, BRAF V600 rises from 10% to 22% across the same stretch, the sigmoid-rectal region is distinct from the rest of the left colon, and the transverse colon clusters with the left.
• The science in detail, 4 of 7, in full; the one-sentence version is on the overview. Two routes in, plus two short cuts. The adenoma-carcinoma sequence runs APC, then KRAS, then 18q and SMAD4, then 17p and TP53, and is mostly left-sided; it starts before any visible lesion, since about 1% of normal crypts in middle age already carry a driver. The serrated pathway starts with BRAF V600E or KRAS in a hyperplastic polyp or sessile serrated lesion, proceeds by CpG island methylation rather than chromosome loss, accounts for about 30% of carcinomas, and is where sporadic mismatch repair deficiency comes from through MLH1 promoter silencing; serrated-class polyps are found in 20 to 40% of average-risk people and are the ones most often missed. Lynch syndrome is the inherited short cut, 3.1% of colorectal cancer probands, caused by a germline MLH1, MSH2, MSH6 or PMS2 variant or a 3' EPCAM deletion. Polymerase proofreading failure is the fourth route: 1.0% of cancers, ultramutated but microsatellite stable, with an excellent prognosis.
• The science in detail, 5 of 7, in full; the one-sentence version is on the overview. Immunity. Mismatch repair deficient tumours carry a mean of 1,782 somatic mutations against 73 in proficient ones, and that is why they respond: 4 of 10 against 0 of 18 in the first pembrolizumab study, 31.1% with single-agent nivolumab in 74 previously treated patients, 16.5 against 8.2 months of progression-free survival with first-line pembrolizumab, and 68% pathological complete response after four weeks of neoadjuvant nivolumab plus ipilimumab. Microsatellite-stable disease has failed every unselected attempt, most clearly IMblaze370, where atezolizumab with or without a MEK inhibitor matched regorafenib in 363 patients. The mechanism is not only low mutation load: TGF-beta-activated stroma excludes T cells and its blockade makes mouse liver metastases checkpoint-sensitive, and tumour-cell WNT activation independently tracks the absence of T cells. Inside the responsive group, biallelic B2M and HLA loss is the documented escape route. Immunoscore, a digital CD3 and CD8 count, is prognostic across the whole disease independently of stage and microsatellite status (hazard ratio 0.20 high against low in 2,681 patients).
• The science in detail, 6 of 7, in full; the one-sentence version is on the overview. Residual disease in blood. Tumour-informed circulating tumour DNA after resection gives recurrence hazard ratios of 7 to 18, after adjuvant chemotherapy 17 to 51, and during surveillance up to 43, with lead times over imaging of about 10 to 16 months; week-4 positivity in 1,039 prospectively followed patients carried a hazard ratio of 10 and 18-month disease-free survival of 38.4% against 90.5%. DYNAMIC randomised 455 patients and cut adjuvant chemotherapy use from 28% to 15% with two-year recurrence-free survival of 93.5% against 92.4%, the only randomised evidence in any solid tumour that a residual disease result can safely change treatment. What is missing is escalation evidence, and the false negatives are real: 16 of 164 ctDNA-negative stage II patients recurred. Plasma sequencing does two other jobs here, watching RAS-mutant clones emerge under EGFR blockade months before imaging, and selecting HER2-amplified patients as accurately as tissue.
• The science in detail, 7 of 7, in full; the one-sentence version is on the overview. Testing, in practice. Mismatch repair immunohistochemistry (MLH1, MSH2, MSH6, PMS2) on every colorectal cancer, with MSI PCR where the stain is equivocal, and BRAF V600E or MLH1 promoter methylation as a reflex where MLH1 is lost; extended RAS across KRAS and NRAS exons 2, 3 and 4 before any EGFR antibody, because a codon 12 and 13 assay misses about one in six KRAS mutations and nearly all NRAS ones; BRAF reported as V600E or non-V600, since the two carry median survivals of 11.4 and 60.7 months and only V600E has a regimen; HER2 scored by the colorectal-specific rule of intense membranous staining in more than 50% of cells, which selects about 5% of RAS wild-type patients; and sequencing rather than staining where POLE ultramutation or an NTRK, RET, ALK or RSPO fusion is possible, since none of those is visible to immunohistochemistry. Primary tumour side belongs on the report alongside the genotype.
• Frequencies marked cBioPortal were computed from the public API on 24 September 2026 on the sequenced sample lists of coadread_tcga_pan_can_atlas_2018 (534 sequenced of 594), coadread_tcga_pub (224 of 276), coadread_dfci_2016 (619), coadread_genentech (72), crc_msk_2017 (1,134), crc_msk_2026 (7,237), crc_eo_2020 (1,516) and crc_sysucc_2022 (1,015), as sample-level counts of non-synonymous mutation, high-level amplification, deep deletion or structural variant. They are not the papers' own percentages, which are quoted alongside.
• Panel content decides what is seen. ACVR2A, BCL9L, RSPO2 and CDX2 read zero in the MSK-IMPACT cohorts because they are not on the panel, not because they are absent from the disease; RSPO3 fusions read 0.44% on a DNA panel against 10% for RSPO2 and RSPO3 together on transcriptome sequencing. Copy-number calls on 20q genes such as PTPRT and GNAS largely reflect the broad chromosome 20q gain of the chromosomally unstable class rather than focal amplification, and the ChangKang copy-number profile calls amplification much more liberally than the others, so its copy-number rows are not used here.
• KRAS and BRAF allele shares are counts of mutation records over all mutation records for that gene in a study, so a tumour with two mutations in one gene counts twice. Study-level mutation rates also differ with cohort design: the prospective population cohort reads BRAF V600E at 17.9% and the metastatic referral cohorts at 7.5 to 7.9%, because sporadic right-sided BRAF-mutant disease is over-represented in resected series and under-represented in metastatic ones.
• Is bowel cancer the same as colorectal cancer? Yes. Bowel cancer is the everyday British name for cancer of the large bowel, which doctors split into colon cancer and rectal cancer and together call colorectal cancer. Cancer of the small bowel and cancer of the anus are different diseases with their own pages. The colon and rectum are counted separately by the world registries and together by most UK and US statistics, which is why the same disease can be the 4th and the 8th commonest cancer in one paragraph and the 3rd in another.
• How common is it? Worldwide, 1,206,011 colon cancers and 778,594 rectal cancers a year with 556,774 and 339,370 deaths (GLOBOCAN 2024), which is about 1.98 million cases and about 896,000 deaths added together. In the UK, 48,213 cases and about 17,700 deaths a year, the 4th commonest cancer and the 2nd commonest cause of cancer death (Cancer Research UK). In the United States, 158,850 cases and 55,230 deaths projected for 2026 (SEER). The lifetime risk in the UK is 1 in 20 for women and 1 in 17 for men.
• Why do I keep reading that it is rising in young people when the overall numbers are falling? Both are true, in different age groups. In the United States, incidence fell 0.9 percent a year over 2013 to 2022 overall, but rose 3 percent a year in adults aged 20 to 49 while falling 2.5 percent a year in the over-65s (American Cancer Society, 2026). In the UK, incidence rates are down 3 percent in the last decade. The same divergence appears in 19 of 36 countries with registry data, and in nine of them rates rose in the young while stable or falling in older adults (Siegel 2019). No cause is established.
• What raises the risk, and how much of it can I change? Cancer Research UK judges 54 percent of UK cases preventable. Eating too little fibre accounts for 28 percent, processed meat 13 percent, overweight and obesity 11 percent, smoking 7 percent, alcohol 6 percent, too little physical activity 5 percent and ionising radiation 2 percent. Type 2 diabetes raises risk 22 to 30 percent and inflammatory bowel disease 70 percent, with a 5 percent risk of cancer after 20 years of colitis. A first-degree relative with the disease more than doubles risk. Age is the largest factor and cannot be changed: 43 percent of UK cases are in people aged 75 and over.
• Who is offered NHS bowel screening, and what does the test do? In England everyone aged 50 to 74 who is registered with a GP is sent a faecal immunochemical test kit through the post every two years; depending on when you turned 50 the first kit arrives at 50, 52 or 54. People aged 75 and over can ask for a kit every two years on the free helpline. The kit looks for blood in a single stool sample you collect at home and post back; results come in about two weeks. Most people are told no further tests are needed. If blood is found you are invited to see a specialist nurse and usually offered a colonoscopy; blood in the poo often turns out to be a fissure or a polyp rather than cancer (NHS; GOV.UK).
• Does screening actually save lives? Yes, and it has been shown in randomised trials rather than inferred. Annual faecal occult blood testing cut 13-year colorectal cancer mortality by 33 percent in Minnesota (Mandel 1993). One flexible sigmoidoscopy offered at 55 to 64 cut incidence by 26 percent and mortality by 30 percent over 17 years, and by 35 and 41 percent among those who attended (Atkin 2017). NordICC, the first randomised trial of screening colonoscopy, cut the 10-year risk of colorectal cancer from 1.20 to 0.98 percent on an invitation basis in a trial where only 42 percent of invitees attended (Bretthauer 2022). Removing the polyps is what does most of the work.
• My GP gave me a poo test rather than referring me. Is that right? It is what NICE asks for. NG12 (1.3.1) tells GPs to offer a quantitative faecal immunochemical test to people with a change in bowel habit, an abdominal mass, iron-deficiency anaemia and several combinations of bleeding, pain and weight loss by age, and to refer on the suspected cancer pathway at 10 micrograms of haemoglobin per gram of faeces or above (1.3.2). Below the threshold the GP should safety net you, and if clinical concern is strong the referral should not wait for the test (1.3.3). A rectal mass is referred without the test. If you were screened recently and have symptoms now, you should still be tested.
• Why does rectal cancer get a magnetic resonance scan when colon cancer does not? Because the rectum sits in a narrow pelvis and the surgeon's plane of dissection runs within millimetres of the tumour. The scan measures how close the tumour comes to that plane, the mesorectal fascia, which becomes the circumferential resection margin in the specimen, and whether tumour has grown into the veins outside the bowel wall. Those two readings decide whether radiotherapy or chemoradiotherapy comes before surgery. In the MERCURY study of 408 patients, magnetic resonance imaging predicted a clear margin with 92 percent specificity, and 327 of the 349 it predicted clear were clear at operation.
• What is watch and wait? In some people the rectal tumour disappears completely after chemoradiotherapy, leaving a flat white scar and a normal scan. Instead of removing the rectum, the team watches with regular examination, endoscopy and magnetic resonance imaging and operates only if the tumour comes back. In 880 such patients in the International Watch and Wait Database, 25.2 percent had a regrowth by two years, almost all of it in the bowel wall and 88 percent of it within two years, and five-year overall survival was 85 percent. In the OPRA trial about half of patients kept their rectum at five years. NICE (NG151 1.3.7) says to tell people who defer surgery that there is a risk of recurrence, that no marker predicts who is safe, and to record the outcome in a national registry.
• My report says the circumferential resection margin is involved. What does that mean? It means tumour was found within 1 mm of the cut surface at the side of the removed rectum. It matters because that is where rectal cancer recurs: in the 52 specimens Quirke examined in 1986, 14 had tumour at the lateral margin and 12 of those recurred in the pelvis. In more than 17,500 patients reviewed in 2008, an involved margin predicted local recurrence, distant metastases and shorter survival, and its predictive power for local recurrence was higher after preoperative radiotherapy than without it. It usually leads to a discussion about more treatment, not to a conclusion about you.
• What happens if the bowel blocks? Where a primary tumour is left in place in metastatic disease, around 20 in 100 people develop obstruction, perforation, bleeding or pain needing surgery (NICE NG151). For acute obstruction on the left side, NICE offers either a stent through the blockage followed by planned surgery a few weeks later, or emergency surgery, when cure is still the aim; for palliative treatment a stent is the option to consider. In the CReST trial, stenting relieved the obstruction in 82.4 percent and reduced the proportion of people left with a stoma from 67.9 to 47.5 percent, with no difference in 30-day mortality, length of stay, recurrence at three years or survival.
• The cancer has spread to my liver. Is that the end? Not necessarily. Colorectal cancer that has spread only to the liver, and sometimes only to the lung, is treated with the aim of cure: resection or ablation of the metastases, sometimes at the same operation as the primary, with chemotherapy before and after (NICE NG151 1.5.15 to 1.5.19). In the EORTC 40983 trial, perioperative chemotherapy added about 9 percentage points to three-year progression-free survival in patients who had their liver metastases resected. Where the metastases are not removable at first, chemotherapy with an EGFR antibody raised the proportion that became operable from 32 to 60 percent in the CELIM trial. In England, around 45 percent of people whose liver metastases were operated on survived five years or more (Morris 2010).
• What is the outlook? Averages hide a wide range and predate the newest drugs. In the United States 65.4 percent of people are alive at five years, which rises to 91.3 percent for the 34 percent found while the cancer is still confined to the bowel and falls to 16.9 percent for distant disease (SEER). In England, around 90 percent survive five years with stage 1, around 85 percent with stage 2, 65 percent with stage 3 and around 10 percent with stage 4, and almost 55 percent of everyone survives ten years (Cancer Research UK). Survival in the UK has more than doubled since the 1970s. These are population figures, not a personal prognosis.

The same record for scripts and assistants; checked 2026-09-24. Data CC BY-NC 4.0, attribute “Data from OnCo (onco.cc)”.