Colon Cancer and Its Surgical Treatment: An Evidence-Based Review (Part 1)

Evidence Grading Key

Two grading systems are used in parallel throughout.

Recommendation strength follows GRADE: Strong (benefits clearly outweigh harms for nearly all patients) or Conditional/Weak (balance is close, values-sensitive, or evidence is uncertain).

Three organising claims of this review:

1. Surgery remains the only reliably curative modality for non-metastatic colon cancer, and the quality of that surgery — R0 margins, intact mesocolic envelope, adequate nodal harvest — is a modifiable determinant of outcome. (Level I–III)
2. The extent-of-resection debate (CME/CVL vs conventional) is not settled by randomised evidence. Observational data favour CME; randomised data show consistent gains in surgical-quality surrogates but have not yet demonstrated a statistically robust survival advantage. (Level II–III, low GRADE certainty)
3. The most consequential recent change in colon cancer management is not surgical technique but biological selection — MMR status, ctDNA, and neoadjuvant immunotherapy are beginning to determine whether, when, and how much surgery is required. (Level II, rapidly evolving)

Executive Summary

The current standard of surgical care

For non-metastatic colon adenocarcinoma, the standard operation is a segmental oncologic colectomy with en bloc removal of the tumour-bearing segment, its mesocolon, and the regional lymphatic basin, achieving an R0 resection. The extent of resection is determined by the lymphovascular territory of the tumour, not by the tumour's linear dimensions. Proximal ligation of the named feeding artery, ≥5 cm longitudinal bowel margins in most cases, an intact mesocolic fascial envelope, and a nodal harvest of at least 12 nodes constitute the minimum acceptable oncologic product.

Minimally invasive surgery is the default access route for most elective colon cancer resections in appropriately experienced hands. This position rests on unusually strong evidence — multiple large randomised trials (COST, COLOR, CLASICC, ALCCaS, JCOG0404) with long-term follow-up demonstrated equivalent oncologic outcomes with faster recovery. (Level I, ⊕⊕⊕⊕, Strong)

Which patients benefit from surgery

Stage-specific summary

• Stage I: Resection alone. No adjuvant therapy. Excellent prognosis (5-year OS typically >90%).
• Stage II: Resection alone for standard-risk disease. Adjuvant fluoropyrimidine is considered but not mandated in high-risk stage II (T4, obstruction, perforation, LVI, PNI, poor differentiation, <12 nodes harvested). dMMR stage II tumours should not receive single-agent fluoropyrimidine — it is ineffective and possibly harmful in this subset. (Level II)
• Stage III: Resection plus adjuvant oxaliplatin-based therapy. Duration is risk-stratified following the IDEA collaboration: 3 months of CAPOX is a reasonable standard for low-risk (T1–3 N1) disease, while 6 months (usually FOLFOX) is preferred for high-risk (T4 and/or N2). (Level I for the duration question)
• Stage III dMMR: The phase 3 ATOMIC trial (NCT02912559) reported that adding atezolizumab to mFOLFOX6 raised 3-year DFS from 76.2% to 86.3% (HR 0.50, 95% CI 0.35–0.73, P<0.001) at 40.9 months' median follow-up, in 712 patients. Overall survival remains immature. This is now reflected in NCCN guidance and represents the first positive phase 3 adjuvant immunotherapy trial in colon cancer. (Level II, ⊕⊕⊕⊝)
• Stage IV: Molecularly directed systemic therapy is the backbone. Surgery is reserved for resectable oligometastatic disease, symptomatic primaries, and complete cytoreduction of limited peritoneal disease.

The major surgical controversies

The most promising emerging strategies

1. Neoadjuvant immunotherapy in dMMR/MSI-H colon cancer. NICHE-2 (n=115, nivolumab + ipilimumab) reported a 99% pathological response rate with 68% pathological complete response, and 3-year DFS of 100% at 36.6 months' median follow-up. This is the single most striking efficacy signal in modern colon cancer. It is single-arm phase 2 evidence in a favourable-prognosis subgroup — but the effect size is large enough that organ-preservation trials are justified. (Level III design, extraordinary effect size)
2. ctDNA/MRD-adapted therapy, once assay sensitivity and optimal sampling timing are resolved.
3. Molecularly targeted first-line therapy in poor-prognosis subtypes. BREAKWATER (NCT04607421) established encorafenib + cetuximab + mFOLFOX6 as a new first-line standard for BRAF V600E mCRC, with median OS reported at 30.3 months versus 15.1 months for standard care.
4. Fluorescence and image-guided surgery for perfusion assessment and lymphatic mapping — plausible, widely adopted, but with a weaker outcome evidence base than adoption rates suggest.

Methodology

Approach

This review followed a PRISMA-informed rather than PRISMA-compliant process. Sources were identified through targeted retrieval of: (a) current guideline documents; (b) landmark randomised trials known to define practice in each domain; (c) recent systematic reviews and meta-analyses (2018–2026) addressing each controversy; and (d) recently reported trials that post-date the most recent guideline revisions.

Source hierarchy applied

Where sources conflicted, precedence was given in the following order:

1. Adequately powered, low-risk-of-bias randomised trials with mature follow-up
2. Meta-analyses restricted to randomised evidence
3. Current guideline recommendations (NCCN, ESMO, ASCRS, ESCP/EAES)
4. Meta-analyses of observational data with explicit certainty grading
5. Large registry or population-based cohorts
6. Single-institution series and expert consensus

Case reports and small uncontrolled series were used only to illustrate technical points, never to support an efficacy claim.

Handling of known biases

• Surgical trials cannot be blinded. Outcome assessment bias is therefore intrinsic; this is noted wherever a subjective endpoint (e.g. complication grading, quality of life) is reported.
• Learning-curve confounding is pervasive in the CME and robotic literature: early-adopter centres are typically high-volume specialist units, so technique and centre quality are entangled. Observational comparisons of CME vs conventional surgery are especially vulnerable.
• Industry sponsorship is flagged where relevant. Robotic platform comparisons frequently involve investigators with device-manufacturer relationships. Among the trials discussed here, ATOMIC was funded by the National Cancer Institute with Genentech support; NICHE-2 was funded by Bristol Myers Squibb; BREAKWATER was funded by Pfizer and collaborators. This does not invalidate their results but is relevant to interpretation of secondary endpoints and framing.
• Publication bias is likely to inflate apparent CME benefit — negative single-centre surgical series are systematically under-published.
• Heterogeneity in surgical definition is the central methodological problem in the CME literature: "CME" has been operationalised inconsistently across studies, which limits the interpretability of any pooled estimate.

Causal language

Observational findings are reported as associations. Causal language is reserved for randomised evidence. Where a meta-analysis of observational data reports a hazard ratio, that ratio is reported alongside its certainty grading and the residual-confounding caveat.

Epidemiology and Disease Burden

Global burden

Colorectal cancer is the third most commonly diagnosed cancer worldwide and the second leading cause of cancer death. GLOBOCAN 2022 estimates place the global burden at approximately 1.93 million new colorectal cancer cases and 904,000 deaths, of which colon (as distinct from rectal and anal) primaries constitute roughly 59%. Eastern Asia carries the largest absolute case numbers; Asia as a whole accounted for approximately half of global cases in 2022.

Incidence correlates strongly with Human Development Index — historically a disease of high-income Western populations, colorectal cancer incidence is now rising in transitioning economies as dietary, activity, and adiposity patterns shift.

The early-onset shift

The most epidemiologically significant recent development is the rise in colorectal cancer diagnosed before age 50. This increase follows a birth-cohort pattern beginning with individuals born in the 1960s, which argues against screening artefact or genetic drift as explanations and points toward environmental or lifestyle exposures acting from early life.

Surgical relevance: early-onset patients are more often diagnosed at advanced stage, more often have left-sided/rectosigmoid primaries, and have decades of survivorship ahead of them — which raises the stakes on functional outcome, fertility preservation counselling, hereditary syndrome testing, and avoidance of unnecessary permanent stomas.

Risk factors

Economic burden

Colorectal cancer is among the most costly malignancies to health systems, driven by the combination of high incidence, surgical intervention in the majority of patients, prolonged systemic therapy in advanced disease, and increasingly expensive targeted and immunotherapeutic agents. Costs are front-loaded (diagnosis and surgery) and back-loaded (terminal care), with a comparatively cheaper middle period. Robotic platform adoption adds capital and per-case disposable costs that are not currently offset by demonstrated oncologic gain — a point taken up in §9.

Surgical Anatomy of the Colon

Oncologic colectomy is a vascular and lymphatic operation performed on an embryological plane. The anatomy below is presented in that frame rather than as descriptive topography.

Embryological basis

The colon derives from two embryological segments with distinct blood supplies and lymphatic drainage:

• Midgut → caecum, ascending colon, hepatic flexure, proximal two-thirds of transverse colon → superior mesenteric artery (SMA)
• Hindgut → distal transverse colon, splenic flexure, descending colon, sigmoid, upper rectum → inferior mesenteric artery (IMA)

The watershed lies near the distal transverse colon, which is why splenic flexure lesions are anatomically awkward: their lymphatic drainage may follow the middle colic, left colic, or both. This is the single most important reason splenic flexure cancer has no universally agreed standard operation.

The mesocolic envelope

The mesocolon is invested in a visceral fascial layer that separates it from the retroperitoneal (parietal) fascia — the colonic analogue of the mesorectal fascia. Dissection in this avascular embryological plane:

• keeps the lymphatic package intact and prevents spillage of tumour-bearing lymph and cells
• protects the retroperitoneal structures beneath (ureter, gonadal vessels, duodenum, pancreas)
• yields a specimen whose mesocolic integrity can be graded pathologically

This plane is the anatomical foundation of the CME concept (§8).

Arterial supply and the ligation decision

The venous problem

The gastrocolic trunk of Henle — the confluence of the right gastroepiploic, superior right colic, and anterior superior pancreaticoduodenal veins draining into the superior mesenteric vein — is the most dangerous structure in right-sided CME. Its branching pattern is highly variable and its walls are thin. Most catastrophic bleeding in CME right hemicolectomy originates here. This anatomical fact is the primary technical argument against unselective adoption of CME.

Lymphatic drainage and nodal stations

Japanese nomenclature (widely used in the CME/D3 literature) divides nodes into:

• D1 — pericolic/epicolic: along the bowel wall and marginal artery
• D2 — intermediate: along the named feeding artery
• D3 — main/central/apical: at the origin of the feeding vessel from the SMA or IMA

Standard Western oncologic colectomy reliably clears D1–D2. CME with central vascular ligation is functionally a D3 lymphadenectomy. The oncologic hypothesis is that D3 nodes harbour disease in a minority of patients (commonly cited at roughly 1–5% for right-sided tumours, rising with T stage) and that removing them converts a fraction of incompletely resected patients into cured ones.

The counter-argument is arithmetic: if D3 node positivity is ~2–4% and only a fraction of those patients are curable by nodal clearance alone (most having systemic disease), the maximum achievable absolute survival benefit is small — and must be weighed against a small but real risk of SMV injury, chylous ascites, and autonomic denervation.

Structures at risk

TNM Staging and Its Surgical Consequences

AJCC/UICC 8th edition (abbreviated)

* Ranges are indicative from published survival series and registry data; they vary substantially by era, region, and treatment received. They should not be used for individual prognostication.

Stage IV survival is the least meaningful summary statistic in oncology. A patient with three resectable liver metastases and MSI-H disease and a patient with M1c peritoneal carcinomatosis and BRAF V600E mutation share a stage designation and essentially nothing else. Stage IV must be subdivided by resectability, metastatic site, tumour burden, and molecular profile before any prognostic statement is meaningful.

How stage drives the operation

Critically, stage does not usually change which colectomy is performed — the resection is dictated by tumour location and vascular territory. Stage changes:

The "high-risk stage II" construct

High-risk stage II is a pragmatic clinical category, not a biological entity. It bundles features of very different mechanistic significance:

The nodal-yield criterion deserves emphasis: it is the only "high-risk feature" that is a failure of the healthcare process rather than a property of the cancer. A stage II patient labelled high-risk for having 8 nodes examined may simply have had a suboptimal operation or a rushed pathological assessment. This is a quality problem being managed with chemotherapy.

dMMR overrides: approximately 15–20% of stage II colon cancers are dMMR. These have a favourable prognosis and do not benefit from single-agent 5-FU, which should not be given on the basis of high-risk features alone in this subgroup. (Level II, Strong)

Principles of Oncologic Colon Surgery

The six non-negotiables

The nodal yield question — what it actually measures

Higher nodal yield associates with better survival. Three explanations compete and are not mutually exclusive:

1. Stage migration (Will Rogers phenomenon): more nodes examined → more occult positives found → more accurate stage assignment → better stage-specific survival in both groups, without any patient benefiting.
2. Surgical quality proxy: surgeons who harvest more nodes are performing more complete resections generally.
3. Therapeutic effect: removing occult nodal disease cures some patients.

The distinction matters enormously for policy. If (1) dominates, mandating higher nodal counts improves statistics, not patients. If (3) dominates, extended lymphadenectomy should be universal. The truth is almost certainly a mixture, weighted toward (1) and (2) — which is precisely why the CME debate remains unresolved.

Confounders on nodal yield that have nothing to do with surgical technique: patient age (yield falls with age), BMI, tumour location (right-sided yields exceed left-sided), immune response to the tumour (dMMR tumours yield more nodes), specimen handling technique, pathologist diligence, and fat-clearing protocols.

En bloc resection in T4b disease

Rule: do not separate a tumour from an adherent structure. Intraoperative distinction between inflammatory and malignant adhesion is unreliable, and division of a malignant adhesion converts a curable R0 operation into an R1 or R2 with local recurrence rates that are dramatically worse. If the tumour is adherent to small bowel, bladder dome, abdominal wall, uterus, or duodenum, the adherent portion goes with the specimen.

This principle is one of the strongest in surgical oncology despite resting on observational evidence (Level III, ⊕⊕⊝⊝, but Strong recommendation on the basis of magnitude and biological plausibility).

Specimen quality grading

Pathological assessment of the mesocolic plane — analogous to Quirke grading of the mesorectum:

Achievement of the mesocolic plane associates with improved survival in observational series. Because plane quality is also a marker of surgeon skill, case selection, and centre volume, the causal contribution of the plane itself remains uncertain.

Surgical Procedures by Tumour Location

Master table

The high tie vs low tie debate

For left-sided and sigmoid tumours, ligation of the IMA at its aortic origin ("high tie") versus distal to the left colic branch ("low tie"):

Arguments for high tie: maximises apical nodal clearance; improves reach for a tension-free anastomosis; conceptually consistent with CVL.

Arguments for low tie: preserves left colic perfusion to the proximal limb (reducing ischaemic leak risk); avoids injury to the superior hypogastric plexus at the aortic bifurcation, thereby reducing sexual and urinary dysfunction.

Evidence: randomised and observational data have consistently failed to demonstrate a survival advantage for high tie. Apical node positivity is low. Meta-analyses generally show no significant difference in leak rates or oncologic outcomes. The autonomic-nerve argument is the strongest differentiator, particularly in younger patients for whom sexual function is a major survivorship concern.

(Level II, ⊕⊕⊕⊝ for equivalence of oncologic outcome; Conditional recommendation favouring low tie where anastomotic reach permits, particularly in young patients.)

Expected complication profiles

Leak rates are indicative ranges from published series and registries; they vary with emergency status, patient factors, and centre. Right-sided (ileocolic) anastomoses leak less frequently than left-sided (colocolic/colorectal) anastomoses — but right-sided leaks, when they occur, are frequently more physiologically severe because of the volume and enzymatic content of small bowel effluent.

Complete Mesocolic Excision and Central Vascular Ligation

This is the central surgical controversy in colon cancer and warrants extended treatment.

Definitions

• CME (Complete Mesocolic Excision): sharp dissection in the embryological plane between the visceral mesocolic fascia and the parietal retroperitoneal fascia, delivering the tumour and its entire mesocolic lymphovascular package within an intact fascial envelope.
• CVL (Central Vascular Ligation): ligation of the feeding artery and its accompanying vein at their origin from the SMA/SMV or aorta/IMA, clearing the D3/apical nodal station.

CME and CVL are conceptually separable — one can perform an excellent plane dissection with non-central ligation, or a central ligation with a torn mesocolon — but they are almost always reported together.

The definitional inconsistency across studies is the field's central methodological weakness. Some series define CME radiologically, some by specimen photography, some by declared surgical intent, some by nodal yield threshold. Pooled estimates across these definitions should be read with corresponding scepticism.

Oncologic rationale

Analogy to TME in rectal cancer: intact fascial envelope → no lymphatic spillage → fewer residual tumour-bearing nodes and cells → lower local recurrence. Plus removal of D3 nodes that a conventional operation leaves behind.

The analogy is imperfect in an important way. TME solved a problem with an extraordinarily high baseline failure rate — local recurrence in rectal cancer before TME was often 20–30%. Local recurrence after conventional colectomy is already low (typically <5–10% in modern series). The headroom for improvement is therefore an order of magnitude smaller, which means a much larger trial is needed to detect a real effect, and that any true effect is necessarily modest in absolute terms.

What the evidence actually shows

Consistent findings (higher certainty)

Contested findings (lower certainty)

The EAES rapid guideline evidence synthesis, conducted with SAGES and ESCP participation (Tzanis et al., Surg Endosc 2025), included 13 studies of which 3 were randomised trials, and applied GRADE formally. It found:

• Overall survival: HR 0.67 (95% CI 0.48–0.93) favouring CME — low certainty
• Disease-free survival: HR 0.78 (95% CI 0.63–0.96) favouring CME — low certainty
• No difference in 30-day mortality, major morbidity, or major blood loss

The authors explicitly attributed the low certainty rating to the predominance of observational data with substantial risk of bias, and called for high-quality randomised trials and technique standardisation.

Other syntheses reach compatible but differently framed conclusions. A multivariate meta-analysis using restricted mean survival time estimation (10 studies, 3,665 patients) found that at 60 months, stage I–III CME patients lived on average 2.5 months longer (95% CI 1.1–4.1) — an effect that is statistically detectable but clinically modest. A 2026 meta-analysis restricted to observational cohorts found the trend toward superior 5-year OS and DFS with CME did not reach statistical significance, and emphasised that the absence of randomised trials prevents attributing the observed differences to technique rather than latent confounding.

The randomised evidence

The randomised base is thin and has not resolved the question:

• RELARC (China): laparoscopic CME with CVL versus D2 dissection. Short-term results showed acceptable safety with more vascular injury in the CME arm. Longer-term disease-free survival numerically favoured CME but did not achieve conventional statistical significance.
• CoME-in (NCT04871399, Italian Society of Surgical Oncology): multicentre randomised superiority trial with 3-year DFS as primary endpoint. Interim analysis published 2023; definitive survival data awaited.
• Several smaller single-centre randomised studies have reported feasibility and short-term safety rather than survival.

No adequately powered randomised trial has yet demonstrated a statistically significant overall survival advantage for CME.

Interpretation

The honest reading:

1. CME reliably improves the surgical product (nodal yield, plane quality). This is not in serious dispute.
2. Whether the improved product translates into improved survival is unproven. The observational signal is consistent but the confounding structure — CME performed by expert surgeons in high-volume centres on selected patients — is exactly the pattern that produces spurious associations.
3. The plausible true effect size is small. Given the low baseline local recurrence rate and the low D3 node positivity rate, an absolute survival gain in the range of 1–4% is the realistic upper bound. Detecting that reliably requires trials far larger than any conducted.
4. The risk profile is not symmetric. Pooled morbidity is equivalent in the hands of the surgeons who published these series. Henle's trunk injury during unsupervised early-experience CME is not a theoretical concern.

Where CME is most defensible

(Conditional recommendations, ⊕⊕⊝⊝)

• Right-sided tumours with radiologically bulky or suspicious central nodes
• cT4 or high-T-stage tumours where the nodal risk is greatest
• Centres with established CME volume, standardised technique, and pathological plane grading feedback
• As part of prospective trial enrolment

Where caution is warranted

• Low-volume settings or early learning curve. The learning curve is real and steep; published estimates for laparoscopic CME competence commonly fall in the range of tens of cases, with vascular complications concentrated early.
• Small, low-stage, node-negative-appearing tumours, where the theoretical benefit is smallest.
• Frail patients in whom a longer operation and a small risk of major venous injury are poorly tolerated.
• Splenic flexure and transverse tumours, where the vascular anatomy is most variable and the dissection least standardised.

Bottom line: CME should be understood as a defensible standard of technical excellence with an unproven survival dividend, not as an established superior operation. Presenting it as automatically superior overstates the evidence; dismissing it ignores a consistent observational signal and a coherent oncologic rationale.

Open vs Laparoscopic vs Robotic Colectomy

The settled question: laparoscopic vs open

This is one of the best-answered questions in surgical oncology. Multiple large randomised trials with mature follow-up — COST (North America), COLOR (Europe), CLASICC (UK), ALCCaS (Australasia), and JCOG0404 (Japan) — established that laparoscopic colectomy for colon cancer produces oncologic outcomes equivalent to open surgery with meaningfully faster recovery.

(Level I, ⊕⊕⊕⊕, Strong recommendation: laparoscopy is the default for elective colon cancer resection where expertise exists.)

Key qualifications:

• The CLASICC trial raised early concerns about circumferential margin positivity in rectal cancer that did not apply to colon cancer, and were subsequently substantially resolved.
• These trials were conducted by credentialled surgeons meeting minimum case volumes. The results do not automatically transfer to a surgeon performing occasional laparoscopic colectomy.
• Conversion to open surgery is not a failure. In several analyses, converted patients have worse outcomes than either completed-laparoscopic or planned-open patients — but this reflects the difficulty of the case that prompted conversion. Early, elective conversion is a mark of judgement; late, reactive conversion after prolonged struggle is not.

The unsettled question: robotic vs laparoscopic

What randomised evidence shows

Meta-analyses restricted to randomised trials in colorectal cancer generally report, for robotic versus laparoscopic surgery:

• Longer operative time — consistently, typically 20–40 minutes
• Lower conversion rate — consistently favouring robotic
• Less estimated blood loss — modest, of uncertain clinical significance
• Similar or slightly lower complication rates — direction favouring robotic in some pooled analyses, not in others
• No difference in lymph node yield
• No demonstrated difference in overall or disease-free survival
• Higher cost — consistently, driven by capital, disposables, and theatre time

A 2025 meta-analysis of robotic versus laparoscopic CME right colectomy (7 studies, 733 patients) found no significant difference in postoperative morbidity (OR 0.95, 95% CI 0.66–1.37, p=0.79). A large synthesis of 46 studies and 36,868 patients (predominantly non-randomised) confirmed the pattern of longer operative times and higher costs with robotic surgery against modest perioperative advantages.

The multicentre randomised ESSIMIC trial (520 patients randomised 2017–2019) has reported interim short-term outcomes for colonic resections; mature oncologic outcomes are pending.

Interpretation

Robotic colectomy is a platform choice, not an oncologic intervention. The current evidence supports it as at least equivalent to laparoscopy with a modest reduction in conversion — which matters most in the situations where laparoscopy is hardest (obesity, bulky tumours, difficult CME, hostile abdomen). It does not support claims of improved cancer outcomes.

Two honest caveats about the literature:

1. Comparison-era bias. Robotic series are newer; laparoscopic comparators often span earlier eras with less mature ERAS pathways.
2. Investigator relationships. Device-manufacturer relationships are common among robotic surgery investigators. This is disclosed variably and warrants attention when reading enthusiastic single-centre series.

(Level II, ⊕⊕⊕⊝ for perioperative equivalence/modest advantage; ⊕⊝⊝⊝ for any oncologic advantage — no such advantage is established.)

When open surgery remains the right choice

Open surgery is not obsolete. It is indicated for:

Comparative summary matrix

Selected Landmark Trials

Key Points

1. Surgery is the only reliably curative modality for non-metastatic colon cancer. The operation is defined by the tumour's vascular and lymphatic territory, not its size.
2. R0 resection, intact mesocolic plane, ≥12 nodes, and avoidance of tumour rupture are the non-negotiable quality standards. (Level I–III)
3. Laparoscopy is the evidence-based default access route. This is settled. (Level I)
4. Robotic surgery reduces conversion and blood loss modestly at higher cost, with no demonstrated oncologic advantage. (Level II)
5. CME/CVL reliably improves specimen quality and nodal yield. Its survival benefit is supported only by low-certainty observational evidence (OS HR ≈0.67, GRADE low) and remains unconfirmed by randomised trials.
6. Do not separate a tumour from an adherent organ. En bloc resection or nothing.
7. "High-risk stage II" bundles biologically heterogeneous features, one of which (<12 nodes) is a healthcare quality failure rather than a tumour property. dMMR status overrides the whole construct for fluoropyrimidine decisions.
8. The frontier is no longer technique but biological selection — MMR status, ctDNA, and neoadjuvant immunotherapy are beginning to determine whether and how much surgery a patient needs.
9. ctDNA is prognostically powerful and not yet clinically actionable for adjuvant de-escalation: DYNAMIC-III did not meet non-inferiority.
10. Every statement above describes populations, not patients. Individual management belongs to a multidisciplinary team.

Abbreviations

• CME, complete mesocolic excision
• CVL, central vascular ligation
• CRS, cytoreductive surgery
• ctDNA, circulating tumour DNA
• DFS, disease-free survival
• dMMR, mismatch repair deficient
• HIPEC, hyperthermic intraperitoneal chemotherapy
• HR, hazard ratio
• ICI, immune checkpoint inhibitor
• IMA, inferior mesenteric artery
• LVI, lymphovascular invasion
• mCRC, metastatic colorectal cancer
• MRD, molecular residual disease
• MSI-H, microsatellite instability-high
• MSS, microsatellite stable
• ORR, objective response rate
• OS, overall survival
• PCI, peritoneal cancer index
• pCR, pathological complete response
• PFS, progression-free survival
• PNI, perineural invasion
• pMMR, mismatch repair proficient
• RFS, recurrence-free survival
• SEMS, self-expanding metal stent
• SMA/SMV, superior mesenteric artery/vein
• TME, total mesorectal excision

1. Bray, F., Laversanne, M., Sung, H., Ferlay, J., Siegel, R. L., Soerjomataram, I., & Jemal, A. (2024). Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: A Cancer Journal for Clinicians, 74(3), 229–263.
2. Chalabi, M., Verschoor, Y. L., Tan, P. B., et al. (2024). Neoadjuvant immunotherapy in locally advanced mismatch repair–deficient colon cancer. New England Journal of Medicine, 390(21), 1949–1958.
3. Chalabi, M., et al. (2024). Neoadjuvant immunotherapy in locally advanced MMR-deficient colon cancer: 3-year disease-free survival from NICHE-2 [Abstract LBA24]. Annals of Oncology, 35(Suppl 2).
4. de Gooyer, P., et al. (2024). Neoadjuvant nivolumab and relatlimab in locally advanced MMR-deficient colon cancer: A phase 2 trial. Nature Medicine, 30, 2968–2976.
5. Elez, E., Yoshino, T., Shen, L., et al. (2025). Encorafenib, cetuximab, and mFOLFOX6 in BRAF-mutated colorectal cancer. New England Journal of Medicine, 392(24), 2425–2437.
6. Degiuli, M., Resendiz Aguilar, A. H., Solej, M., et al. (2023). A randomized phase III trial of complete mesocolic excision compared with conventional surgery for right colon cancer: Interim analysis of the CoME-in trial. Annals of Surgical Oncology. https://doi.org/10.1245/s10434-023-14664-0
7. Morton, D., Seymour, M., Magill, L., et al. (2023). Preoperative chemotherapy for operable colon cancer: Mature results of an international randomized controlled trial (FOxTROT). Journal of Clinical Oncology, 41(8), 1541–1552.
8. Quénet, F., Elias, D., Roca, L., et al. (2021). Cytoreductive surgery plus hyperthermic intraperitoneal chemotherapy versus cytoreductive surgery alone for colorectal peritoneal metastases (PRODIGE 7): A multicentre, randomised, open-label, phase 3 trial. The Lancet Oncology, 22(2), 256–266.
9. Sinicrope, F. A., Ou, F.-S., Arnold, D., et al. (2026). Atezolizumab plus FOLFOX for stage III mismatch repair–deficient colon cancer. New England Journal of Medicine, 394(12), 1115–1126.
10. Tie, J., Wang, Y., Loree, J. M., et al. (2025). Circulating tumor DNA–guided adjuvant therapy in locally advanced colon cancer: The randomized phase 2/3 DYNAMIC-III trial. Nature Medicine, 31, 4291–4300.
11. Tzanis, A. A., Antoniou, S. A., Carrano, F. M., et al. (2025). A systematic review, meta-analysis and GRADE assessment of the evidence on complete mesocolic excision for right-sided colon cancer with SAGES and ESCP participation. Surgical Endoscopy, 39(6), 3466–3473.
12. National Comprehensive Cancer Network. (2026). NCCN clinical practice guidelines in oncology: Colon cancer (Version 2.2026, April 7, 2026). NCCN.