Snapshot. TWO DISEASES IN ONE NAME. COLON = surgery FIRST → adjuvant chemo; NEVER radiotherapy. RECTUM = TME (total mesorectal excision)-based, radiotherapy-relevant, treated NEOADJUVANTLY. The dividing line is ANATOMICAL: >12 cm from the anal verge / above the peritoneal reflection behaves as COLON. Molecular drivers: RAS, BRAF V600E, MMR/MSI (mismatch-repair / microsatellite-instability status), HER2, KRAS G12C, and SIDEDNESS (left = better prognosis + anti-EGFR works; right = worse + anti-EGFR does NOT work).

Colonoscopy + biopsy — height from the anal verge is the whole colon-vs-rectum fork; clear the whole colon (~5% synchronous second primary) · CT chest/abdomen/pelvis (liver + lung are the target organs) · RECTAL MRI if rectal — the decision-maker: mrT substage (T3a/b vs T3c/d), CRM (circumferential resection margin, threatened if ≤1 mm), EMVI (extramural vascular invasion), lateral nodes — "T3" alone is not a stratum · CEA baseline · MMR/MSI — UNIVERSAL, every patient (prognosis, adjuvant decision, IO eligibility, + Lynch syndrome screen) · RAS/BRAF V600E/HER2/KRAS G12C if metastatic (biologic selection; anti-EGFR only if RAS wild-type) · DPYD before any fluoropyrimidine — deficiency causes fatal toxicity.

CEA every 3–6 months · CT chest/abdomen/pelvis annually (more often if high risk) · colonoscopy at 1 year, then by findings. The purpose is to catch RESECTABLE recurrence — oligometastatic relapse can still be cured.

⚠ The surveillance paragraph above is written for a RESECTED patient. A watch-and-wait patient still has his rectum and his tumour bed in him, and CEA 3–6 monthly with an annual CT would miss a regrowth while it is still curable. Organ preservation is legitimate ONLY on a documented clinical complete response (cCR). Assumed from a good-looking scan, it is not organ preservation, it is undertreatment.
① The cCR TRIAD — all three, assessed 8–12 weeks after the end of radiotherapy: DRE (flat soft scar, no palpable mass, nodule or ulcer) · endoscopy (flat white scar ± telangiectasia; no residual ulcer, nodule or mass) · rectal MRI with DWI (diffusion-weighted imaging) (tumour replaced by low-signal fibrosis, no restricted diffusion, no suspicious nodes). A NEAR-complete response is not a failure and not an indication to operate — reassess in 4–8 weeks, a real proportion convert to complete.
② The MMR/MSI status forks the pathway, so settle it first: pMMR (mismatch-repair proficient) → OPRA-style watch-and-wait, i.e. chemoradiation plus chemotherapy then observation; dMMR / MSI-high → the neoadjuvant PD-1 (dostarlimab) organ-preservation route instead — same goal, different treatment.
③ The follow-up is intensive and front-loaded, because the risk is: the large majority of regrowths appear within 2 years and almost all within 3 → DRE + endoscopy every 3 months for the first 2–3 years, then less often · rectal MRI every 6 months · CEA every 3 months · CT chest/abdomen/pelvis for distant disease · about 5 years in total.
④ Regrowth is SALVAGEABLE, and that is the fact the whole strategy rests on: TME (total mesorectal excision) performed for a regrowth does not appear to cost survival (OPRA). The danger is therefore not the regrowth, it is the missed appointment. The patient is trading one operation for years of attendance — counsel him in those terms, and do not offer this to someone who cannot commit to the follow-up.

Autosomal-dominant MISMATCH-REPAIR defect (MLH1/MSH2/MSH6/PMS2, or EPCAM deletion) → dMMR / MSI-high tumours. Two classic types: LYNCH I = site-specific colorectal only; LYNCH II = colorectal + the extracolonic spectrum (endometrial, gastric, ovarian, urothelial, small-bowel, pancreatobiliary) — with the Muir-Torre variant (adds sebaceous skin tumours) and Turcot (adds CNS). (Historical labels — now classified by the GENE, which actually sets the risks: MLH1/MSH2 highest, MSH6/PMS2 lower + endometrial-skewed.)
⚠ RED FLAGS to test: young-onset (<50) · right-sided · dMMR/MSI-high tumour · OR a family cluster of Lynch-spectrum cancers (colorectal · endometrial · gastric · ovarian · urothelial · small-bowel). Universal tumour MMR/MSI already screens it → if dMMR, reflex BRAF V600E + MLH1-methylation (both present = sporadic, NOT Lynch) → then germline multigene panel (MLH1/MSH2/MSH6/PMS2/EPCAM; add MUTYH/APC if polyposis or consanguinity). Lynch changes 4 things: ① the OPERATION — discuss extended (subtotal/total) colectomy for the high metachronous-cancer risk, esp. in a young patient; ② SURVEILLANCE — colonoscopy q1–2 yr + gene/family-directed (endometrial/ovarian in women, upper GI if a gastric-cancer family, urinary); ③ prophylactic ASPIRIN lowers CRC risk (CAPP2); ④ CASCADE testing of relatives (children's predictive testing deferred to adulthood — Lynch cancers are adult-onset). dMMR also = IO-sensitive if advanced + no single-agent 5-FU at stage II.