SPECIAL ARTICLE

Localised colon cancer: ESMO Clinical Practice Guidelines for diagnosis,

treatment and follow-upy
G. Argilés1, J. Tabernero2, R. Labianca3, D. Hochhauser4, R. Salazar5, T. Iveson6, P. Laurent-Puig7,8,9, P. Quirke10, T. Yoshino11,
J. Taieb7,8,9,12, E. Martinelli13 & D. Arnold14, on behalf of the ESMO Guidelines Committee*
1
Department of Medical Oncology, Vall d’Hebron University Hospital and Institute of Oncology (VHIO), Universitat Autònoma de Barcelona; 2Department of Medical
Oncology, Vall d’Hebron University Hospital and Institute of Oncology (VHIO), UVic-UCC, IOB-Quiron, Barcelona, Spain; 3Department Oncology, Ospedale Papa
Giovanni XXIII, Bergamo, Italy; 4UCL Cancer Institute, London, UK; 5Department of Medical Oncology, Catalan Institute of Oncology, Oncobell Program (IDIBELL),
CIBERONC, Hospitalet de Llobregat, Barcelona, Spain; 6University Hospital Southampton, NHS Foundation Trust, Southampton, UK; 7Assitance Publique-Hôpitaux de
Paris AP-HP Paris Centre, Paris; 8Paris Cancer Institute CARPEM, Centre de Recherche des Cordeliers, Paris Sorbonne University, Paris University, Paris; 9INSERM, CNRS,
Paris, France; 10Pathology and Data Analytics, School of Medicine, University of Leeds, Leeds, UK; 11National Cancer Centre Hospital East, Kashiwa, Japan;
12
Department of Gastroenterology and GI Oncology, Georges Pompidou European Hospital, Paris Descartes University, Paris, France; 13Università degli Studi della
Campania Luigi Vanvitelli, Department of Precision Medicine, Naples, Italy; 14Asklepios Tumorzentrum Hamburg, AK Altona, Hamburg, Germany

Available online 20 July 2020

Key words: localised colon, Clinical Practice Guidelines, diagnosis, treatment and follow-up

INTRODUCTION characteristics and genetically determined factors.

Screening tests are modulated according to the individual
Incidence and epidemiology
probability of developing CRC.7e9 Age is considered the
Colorectal cancer (CRC) is the third most common tumour in major unchangeable risk factor for sporadic colon cancer:
men and the second in women, accounting for 10% of all nearly 70% of patients are >65 years of age and this disease
tumour types worldwide. Incidence is 25% higher in males is rare before the age of 40 years, even though data
and differs greatly between countries. With more than from Western registries show an increased incidence in the
600 000 deaths estimated each year, CRC is the fourth most 40e44-year age group.10
common cancer-related cause of death globally.1,2 The Individuals with any of the following are considered at
growing incidence in some countries reflects a modification high risk of colon cancer and must be actively screened and
in lifestyle and its consequences related with ‘Westernisa- in case of inherited syndromes, also referred for genetic
tion’ such as obesity, physical inactivity, alcohol consump- counselling (see ESMO guidelines for hereditary gastroin-
tion, high red meat intake and cigarette smoking.3 Some testinal cancer11):
data suggest a putative role in colon cancer carcinogenesis
for factors that cause imbalances in gut microbiota.4,5 ! a medical history of adenoma, colon cancer, inflamma-
The mortality rate in the European Union is 15e20 out of tory bowel disease (Crohn’s disease and ulcerative
100 000 in males and 9e14 out of 100 000 in females and colitis);
has decreased over time, particularly in females. In affected ! significant family history of CRC or adenoma;
European individuals, 5-year survival ranges from 28.5% to ! an inherited cancer syndrome (2%e5% of all CRC), such
57% in men and from 30.9% to 60% in women, with a as familial adenomatous polyposis coli and its variants
pooled estimation in 23 countries of 46.8% in men and (1%), Lynch-associated syndromes (hereditary non-
48.4% in women.6 polyposis colon cancer) (2%e4%), Turcot, Peutze
The risk of developing colon cancer depends on factors Jeghers and MUTYH-associated polyposis syndrome.
which can be classified into lifestyle or behavioural

SCREENING PRINCIPLES
*Correspondence to: ESMO Guidelines Committee, ESMO Head Office, Via
Ginevra 4, 6900 Lugano, Switzerland CRC arises following progression of normal mucosa to an
E-mail: clinicalguidelines@[Link] (ESMO Guidelines Committee). invasive tumour, passing through different intermediate
y
Approved by the ESMO Guidelines Committee: April 2002, last update May stages of premalignant and invasive malignant lesions; this
2020. This publication supersedes the previously published versiondAnn Oncol. stepwise process facilitates cancer prevention and early
2013;24(suppl 6):vi64-vi72.
0923-7534/© 2020 European Society for Medical Oncology. Published by diagnosis when the tumour is still at an early stage and
Elsevier Ltd. All rights reserved. curable, through screening programmes. For average-risk

Volume 31 - Issue 10 - 2020 [Link] 1291

Annals of Oncology G. Argilés et al.

populations, European and American evidence-based (3.6%)17 with identical or different histological patterns and
guidelines for quality assurance in CRC screening12,13 stages of development. Metachronous primary tumours
should be followed. arise in up to 3% of cases during the 5 years after surgery,
and the incidence increases up to 9% after several decades
Recommendations in long-term survivors, justifying long-term surveillance of
the colon in patients who have already experienced colon
Colonoscopic tests cancer.18
! Colonoscopic techniques, despite being invasive, have
the advantage of being both diagnostic and therapeutic. Diagnostic work-up
! A complete colonoscopy is the recommended method A complete work-up should be carried out to achieve an
for CRC screening in average-risk men and women based accurate histological diagnosis of the primary tumour,
on higher sensitivity and specificity when compared with assess the baseline characteristics of the patient and
other tests14 [II, B]. The optimal age range for testing is determine the extent of the disease (see Table 1).
50e74 years [V, C] with an optimal repetition interval
for a negative test of 10 years [III, C]. Diagnosis of the primary tumour. In the absence of a bowel
! Flexible sigmoidoscopy (FS) carried out every 5e10 years obstruction or massive haemorrhage, which may constitute
may be an alternative for those who refuse colonoscopy indications of an urgent tumour resection, a total colonos-
[II, B]. The combination of this method with a yearly copy is recommended for diagnostic confirmation of colon
faecal occult blood test (FOBT) (see below) is recommen- cancer [I, A]. There are many advantages of endoscopy
ded to reduce the risk of a right colon tumour [III, B]. including determination and marking of the exact tumour
! Other invasive tests including capsule colonoscopy are location and biopsy of the lesion and detection and removal
not recommended for screening [IV]. of (further) synchronous precancerous or cancerous lesions.
Combining the limited left-sided colonoscopy with
computed tomography (CT), colonoscopy is an alternative if
Non-invasive tests full colonoscopy is not feasible [I, A].19 In cases where
complete colonic exploration cannot be carried out before
! Non-colonoscopic tests are recommended in average-
surgery, a complete colonoscopy should be carried out
risk men and women from the age of 50 not already
within 3e6 months [IV, B].
taking part in colonoscopic screening programmes. The
optimal frequency of testing is every year and no later Assessment of patient baseline status and characteristics.
than every three years [I, B]. A colonoscopy must be After colonic tumour diagnosis, clinical examination and
carried out at the earliest convenience when the test laboratory tests must be carried out to provide a correct
results are positive [I, A]. assessment of patient status and characteristics before
! Among the available tests, faecal immunochemical deciding the definitive treatment approach [II, A].
testing (FIT) appears to be superior to high-resolution Besides a comprehensive physical examination20 [IV],
guaiac FOBT with respect to the detection rate and pos- blood tests including complete blood count, coagulation,
itive predictive value for adenomas and cancer [III]. liver and kidney functions tests as well as albumin can
Other novel methods including DNA-based or tests using
other markers (e.g. M2-PK) lack formal comparisons of
their performance, and integration with other assays Table 1. Diagnostic work-up for localised CRC
needs to be monitored.
Local assessment LoE, GoR

Screening for high-risk populations is covered in the Complete colonoscopy I, A

Imaging work-up
ESMO guidelines for hereditary gastrointestinal cancer.11 CT scan:
! Lung V
DIAGNOSIS ! Abdominal I, B
! Pelvic I, B
CT colonography (when complete colonoscopy is not I, A
Symptoms and signs feasible)
Colon cancer arises from the mucosa of the bowel, growing MRI abdominal (to clarify ambiguous lesions or define II, A
pT4b)
both into the lumen and the bowel wall, and/or spreading Laboratory work-up
to adjacent organs. Symptoms are associated with relatively Complete blood count II, A
large tumours and/or advanced disease stages and may not Coagulation II, A
Liver function panel II, A
be specific for colon cancer. Alterations in bowel habit, Kidney function panel II, A
general or localised abdominal pain, weight loss without Albumin III, A
other specific causes, weakness, iron deficiency and CEA III, A
anaemia are the most common symptoms and depend on CEA, carcinoembryonic antigen; CRC, colorectal cancer; CT, computed tomography;
GoR, grade of recommendation; LoE, level of evidence; MRI, magnetic resonance
the location and stage of the primary tumour.15,16 Colon imaging.
cancer can occur with multiple or synchronous lesions

1292 [Link] Volume 31 - Issue 10 - 2020

G. Argilés et al. Annals of Oncology

provide relevant clinical information regarding the patient’s ! When not carried out before or during the surgical pro-
baseline conditions and the existence of cancer-related cedure, a complete colonoscopy should be carried out
complications [II, A]. within 3e6 months following tumour resection [IV, B].
In addition, serum levels of carcinoembryonic antigen ! Comprehensive physical examination and laboratory
(CEA), although not sufficient for colon cancer diagnosis tests including full blood counts, biochemistry and serum
themselves in the absence of a confirmatory tumour biopsy CEA levels must be carried out before decisions on the
(because of low specificity and sensitivity), should be eval- definitive treatment approach [III, A].
uated before surgery and monitored during the follow-up ! CT of the thoracic, abdominal and pelvic cavities with i.v.
period to help the early detection of metastatic disease contrast administration is the preferred radiological
[III, A].21e23 In addition, CEA level determination after colon method for the evaluation of the extent of CRC [II, B].
cancer diagnosis is of particular importance since baseline ! Contrast-enhanced MRI constitutes the reference test for
levels add information in defining prognosis; a post- evaluation of the relationship of locally advanced
operative serum CEA level >5 ng/ml (or even >2.35) sug- tumours with surrounding structures or in defining
gests a worse outcome.21 ambiguous liver lesions [II, A].
Assessment of distant tumour extension. Preoperative
assessment of tumour extension is required to determine MANAGEMENT OF LOCALISED COLONIC TUMOURS
whether the patient should be referred for primary tumour
resection or, in the presence of unresectable distant me- Treatment of adenocarcinomas presenting in adenomas
tastases, systemic therapy. Approximately 20% of newly Complete en bloc endoscopic resection should be carried
diagnosed colon cancers have synchronous metastases, the out whenever the morphological structure of the polyp
most frequently involved organ being the liver (17%), permits.34 Endoscopic resection is sufficient for hyperplastic
followed by peritoneum (5%), lung (5%) and lymph nodes or adenomatous polyps and non-invasive (pTis, i.e. intra-
(3%).24 epithelial or intramucosal) adenocarcinomas35 (see
CT of the thoracic, abdominal and pelvic cavities with Figure 1). For (pT1) invasive carcinomas, the management is
intravenous (i.v.) contrast administration is the preferred determined by the polyp morphology and the presence of
radiological method for the evaluation of the presence of histological features associated with adverse outcome36:
distant metastases of CRC [II, B]. This test allows evaluation
of locoregional tumour extension and its complications (e.g. ! lymphatic or venous invasion;
obstruction, perforation, fistula, abscess).25 Nevertheless, ! grade 3 differentiation;
CT scanning may fail to detect peritoneal metastases, where ! significant (grade >1) tumour budding.37
sensitivity is relatively poor and depends on implant local-
isation and size.26,27 For a pedunculated polyp with a pT1 carcinoma confined
Contrast-enhanced magnetic resonance imaging (MRI) to the head, neck and stalk (Haggitt 1e3) endoscopic
permits better definition of the soft tissues. It constitutes resection with proper follow-up is enough even with the
the reference test when it is necessary to evaluate the presence of submucosal invasion, provided that no other
relationship of locally advanced tumours with surrounding unfavourable factors are present [IV, B].38 However, the
structures or in defining ambiguous liver lesions previously presence of any unfavourable factor in a sessile or flat polyp
detected by CT scan [II, A].28 Likewise, MRI can substitute (Paris classification) with a pT1 carcinoma mandates surgical
for CT scanning in patients with iodine-contrast allergies or resection in patients with average operative risk [IV, B].39
chronic renal insufficiency where the glomerular filtration The goal of surgical resection is complete lesion resection,
rate is <30 ml/min [II, A].29e31 including lymph node removal for optimal risk assessment
Fluorodeoxyglucose positron emission tomography (FDG- [IV, B]. In contrast, finding positive resection margins (<1
PET), with or without integrated CT (positron emission to- mm) constitutes only a risk for local recurrence and can be
mography/CT), does not add significant information to the managed by excision repetition or local surveillance.39
CT scans on preoperative staging of CRC and is not rec- When surgery is not possible due to significant comor-
ommended for routine use in staging of localised CRC bidities, surveillance colonoscopy within 6 months after
beyond assisting in interpretation of ambiguous findings [II, polyp removal is recommended, as well as close oncological
A].32,33 follow-up including CT scan to detect lymph node
recurrences [IV, B].38,39
Recommendations
Management of locally infiltrative colon cancers
! In the absence of indications for urgent tumour resec- Infiltrative colon cancers cannot be resected by colonoscopy
tion, a total colonoscopy is recommended for diagnostic and require surgery, with the goal of wide resection of the
confirmation of colon cancer and to rule out synchro- involved bowel segment and its lymphatic drainage [I, A].
nous tumours. Combining the limited left-sided colonos- The extent of the colonic resection is determined by the
copy with CT colonoscopy is an alternative if full blood supply and distribution of regional lymph nodes. The
colonoscopy is not possible [I, A]. resection should include a segment of colon of at least 5 cm

Volume 31 - Issue 10 - 2020 [Link] 1293

Annals of Oncology G. Argilés et al.

Figure 1. Diagnostic algorithm for localised colon cancer.

on either side of the tumour, but wider margins are often ileorectal anastomosis or segmental resection after intra-
included due to the mandatory ligation of the arterial blood operative colonic lavage are alternatives in selected cases
supply [IV, B]. En bloc colonic and mesentery resection is [III]. Colonic stenting45,46 can be used in expert centres as a
recommended in order to clearly define stage II versus stage bridge to elective surgery, especially in patients with higher
III and to identify and eradicate potential lymph node me- rates of postoperative complication after emergency
tastases; at least 12 lymph nodes should be resected when surgery [>70 years old and/or American Society of Anes-
feasible [IV, B].40 Likewise, en bloc resection of adjacent thesiologists (ASA) >II] [II].
organ-invaded portions must be carried out in case of
pT4b41 [I, B].
During the procedure, a complete assessment of the Recommendations
peritoneal cavity and ovaries should be carried out to
investigate for possible metastasis41 [I, C]. (See ESMO ! En bloc endoscopic resection of the polyp is sufficient for
guidelines for metastatic colorectal cancer for the non-invasive (pTis, i.e. intraepithelial or intramucosal)
management of patients with removed metastases42). adenocarcinomas [IV, B].
Laparoscopic colectomy can be safely carried out for ! The presence of invasive carcinoma (pT1) in a polyp
colon cancer when technical expertise is available in the requires a thorough review with the pathologist and
absence of contraindications in view of reduced morbidity, surgeon. High-risk features mandating surgical resection
improved tolerance and similar oncological outcomes with lymphadenectomy include lymphatic or venous
[I, C].43,44 invasion, grade 3 differentiation, significant (grade >1)
Obstructive CRCs can be treated in one or two stages. and tumour budding [IV, B].
Two-stage procedures can include colostomy followed by ! Laparoscopic colectomy can be safely carried out for
colonic resection or, in the case of bowel perforation, colon cancer when technical expertise is available in
Hartmann’s procedure followed by colostomy closure and the absence of contraindications, in view of reduced
anastomosis. One-stage procedures are preferred when morbidity, improved tolerance and similar oncological
carried out by experienced teams; subtotal colectomy and outcomes [I, C].

1294 [Link] Volume 31 - Issue 10 - 2020

G. Argilés et al. Annals of Oncology

! Obstructive CRCs can be treated in one- or two-stage benefits and complications from the given adjuvant treat-
procedures, as indicated [III, B]. ment (see Figure 2).

Assessment of recurrence risk and expected benefits from

PATHOLOGICAL REPORT adjuvant therapy
Pathological reporting should be carried out at the time of The assessment of risk of recurrence is important in
surgery to precisely define nodal spread of disease and deciding when to recommend systemic adjuvant treatment
extension of the tumour through the bowel wall and on to with the aim of reducing risk of relapse and death. The risk
adjacent structures, as well as to assess biopsies when a of relapse after colon cancer resection is estimated by
suspicion of liver or peritoneal metastases has been identified integrating the clinicopathological features of the tumour
by the surgeon. The standard assessment should include47: with the molecular marker MMR/MSI status.49
TNM staging remains the most relevant histological
! morphological description of the specimen; criteria for risk assessment after surgery of colon cancer.
! surgical procedure carried out; Reported 5-year survival rates after surgical resection alone
! definition of tumour site and size; are 99% for stage I, 68%e83% for stage II and 45%e65% for
! presence or absence of macroscopic tumour perforation; stage III disease.48
! histological type and grade; In addition, for intermediate stage II, further parameters
! extension of tumour into the bowel wall and adjacent need consideration to refine the evaluation of risk given the
organs (T stage); observed variability on prognosis [II]49:
! distance of cancer from resected margins (proximal, Major prognostic parameters for stage II risk assessment
distal and radial); [II]48e50:
! presence or absence of tumour deposits;
! lymphovascular and/or perineural invasion; ! Lymph nodes sampling <12;
! presence of tumour budding37; ! pT4 stage including perforation;
! site and number of removed regional lymph nodes and
their possible infiltration by cancer cells (N stage); Minor prognostic parameters for stage II risk assess-
! involvement of other organs (e.g. peritoneum) if submit- ment49 [II]:
ted either removed or biopsied (M stage);
! mismatch repair (MMR)/microsatellite instability (MSI) ! High grade tumour;
status of the tumour. ! Vascular invasion;
! Lymphatic invasion;
The pathological stage must be reported according to the ! Perineural invasion;
Union for International Cancer Control (UICC) tumour, node, ! Tumour presentation with obstruction;
metastasis (TNM) classification, 8th edition48 (see ! High preoperative CEA levels.
supplementary Table S1, available at [Link]
016/[Link].2020.06.022). In general, it has been established that adjuvant systemic
therapy decreases the risk of death by an absolute 3%e5%
in high-risk stage II colon cancer with single-agent 5-
Recommendation fluorouracil (5-FU) and by 10%e15% in stage III disease
with fluoropyrimidines alone, with a further 4%e5%
! A standard surgical/pathological report should include improvement with oxaliplatin-containing combinations
specimen description and surgical procedure, tumour [I, A].
site and size, macroscopic tumour perforation, histologi- MSI/MMR status is the most validated prognostic
cal type and grade, extension into the bowel wall and molecular marker used in deciding adjuvant therapy next
adjacent organs, distance of cancer from resected mar- to clinical prognostic factors.
gins (proximal, distal and radial), presence or absence Deficient DNA MMR status can be identified by immuno-
of tumour deposits, lymphovascular and/or perineural histochemistry detecting loss of MMR protein expression
invasion, tumour budding, site and number of removed (MLH1, MSH2, MSH6 or PMS2) or by polymerase chain
and involved regional lymph nodes, MMR/MSI status reaction (PCR) assays of MSI status (microsatellite muta-
and involvement of other organs [IV, A]. tions). Determining MSI/MMR status in localised colon
cancer patients has two objectives: to characterise the
prognosis and prediction of adjuvant benefit and determine
potential genetic predisposition.
RISK ASSESSMENT MSI/MMR status determination is important to rule out
Definitive decisions regarding adjuvant treatment indication Lynch syndrome. The presence of MSH2 and/or MSH6 loss
can only be made after discussing in detail the risk/benefit by IHC indicates suspicion of Lynch syndrome, while MLH1
ratios of available options with the patient. To this end, the and PMS2 loss needs to be investigated further by deter-
risk of tumour recurrence must be integrated with expected mining BRAF mutation or hypermethylation of the

Volume 31 - Issue 10 - 2020 [Link] 1295

Annals of Oncology G. Argilés et al.

(EMA) dated 13 March 2020, testing for DPD insufficiency

should be conducted before initiating fluoropyrimidine-
based chemotherapy [III, A]. There are two main ways to
assess DPD functionality: through genotyping the DPYD
gene or through phenotyping DPD function.
Genotyping identifies pathologic polymorphisms in the
DPYD gene, mainly DPYD*2A, c.1679T>G, c.2846A>T or
c.1236G>A.56 In the presence of a heterozygous poly-
morphism, the fluoropyrimidine dose should be reduced by
50%, while with homozygous polymorphisms, fluoropyr-
imidines should not be used due to the high risk of com-
plications [III, A], according to a Dutch cohort observational
trial.57 Phenotyping allows assessment of DPD functionality
by measuring the uracilemia in blood.58e61 For levels >16
ng/ml, the dose should be reduced by 50% and for levels
>150 ng/ml, fluoropyrimidines are contraidicated [III, A].57
In this situation, raltitrexed may be an option for those
patients with high risk of recurrence [V].62
Age is another criterion for risk assessment in the adjuvant
Figure 2. Factors to guide adjuvant decision making. setting although it remains controversial. Analyses from a
Canadian database (N ¼ 2801) in Ontario indicate that pa-
tients in stage III disease between the age of 70 and 79 years
promoter region of hMLH1. The identification of either of received adjuvant treatment in 68% and for patients >80
these alterations suggests with high probability the pres- years in 24%.63 In this retrospective analysis, all age groups
ence of an MLH1 gene somatic acquired alteration rather benefited about the same level. However, the indication for
than Lynch syndrome.11 Besides its implications for Lynch an adjuvant treatment had to be associated with the Charlson
syndrome diagnosis, MSI/MMR status defines, in localised Comorbidity Index, ensuring that only ‘fit’ elderly patients
colon cancer, a subgroup of patients with a better prognosis receive an adjuvant treatment. However, all generalisations
and less expected benefit from chemotherapy.51e55 In from clinical randomised trials are difficult to do, since pa-
particular, MSI/MMR may be useful to identify a small tients >75 years are underrepresented and/or excluded.
(10%e15%) subset of stage II patients who are at a very low On the other hand, the addition of oxaliplatin to any
risk of recurrence and in whom the benefits of fluoropyr- fluoropyrimidine should be used with caution in this pop-
imidines have not been demonstrated and thus adjuvant ulation.63,64 A pooled analysis from 4 randomised trials,
chemotherapy should not be indicated [I, A].51e55 NSABP-C08, XELOXA, X-ACT and AVANT, has shown that in
Nomograms have been developed as tools to standardise all age groups, treatment with oxaliplatin can be consid-
decision making in the adjuvant setting; however, their use ered, if clinically indicated.65 The hazard ratio (HR) for
is not widely implemented.56 overall survival (OS) with oxaliplatin was 0.78 for patients of
70 years or older; however, younger patients experienced a
Assessment of risk of complications from adjuvant greater benefit (HR 0.62) and had a significantly lower rate
treatment of toxicity. Similar data were demonstrated in the NO16968
Administration of an adjuvant treatment should only be trial (XELOX versus bolus 5FU/FA: HR for OS in patients
done by experienced sites, with a good knowledge of 70 years or older: 0.91 (0.66e1.26) versus 0.80 at younger
side-effects and (necessary) dose reduction schedules. patients).64 A similar existing but reduced benefit also
Despite the proven benefit for patients with stage III and II occurred in the analysis of the ACCENT database.66
disease, the (relative) counterindications have to be
considered: e.g. Eastern Cooperative Oncology Group
(ECOG) performance status >2, uncontrolled infection, Use of personalised medicine in localised colon cancer/
severe liver and renal dysfunction and heart failure [New biomarkers for risk assessment
York Heart Association (NYHA) III and IV]. Furthermore, Besides MSI status, other genetic markers, e.g. RAS and
other life-prognosis determining comorbidities have to be BRAF mutations are not recommended for the routine
taken into account. assessment of risk of recurrence in non-metastatic patients,
Dihydropyrimidine dehydrogenase (DPD) is the main based on their lack of utility in the adjuvant decision-making
enzyme involved in fluoropyrimidine metabolism. Approxi- process.67 However, other biomarkers such as gene signa-
mately 3%e5% of patients have deficiencies of DPD func- tures, Immunoscore and postoperative circulating tumour
tion due to genetic polymorphisms leading to increased DNA (ctDNA) have demonstrated some benefit in deter-
fluoropyrimidine toxicity, that can be lethal.57 Based on the mining the risk of recurrence and can be considered in
recommendation of the Pharmacovigilance Risk Assessment addition to pathological features and MSI status to further
Committee (PRAC) of the European Medicines Agency tailor the adjuvant decision making in difficult cases.68e71

1296 [Link] Volume 31 - Issue 10 - 2020

G. Argilés et al. Annals of Oncology

Gene signatures have emerged as potential candidates be taken into consideration for refining the risk assess-
for prognostic stratification in locoregional disease. At the ment on stage II tumours [III, A].
time of writing, only Oncotype DX68 and GeneFx Colon69 ! Patient age alone has no predictive value for or against
have been validated in multivariate analysis of indepen- the indication to an adjuvant treatment and must be
dent prospective randomised cohorts of stage II colon considered in the context of (potential) benefit, underly-
cancer with formalin-fixed paraffin-embedded (FFPE) ing risk for relapse, life expectancy in relation to (biolog-
tumour samples. Although routine clinical utility is not ical) age and comorbidities. However, it can be
warranted due to lack of predictive value for chemotherapy generalised that benefits of treatment with both, fluoro-
benefit and the small prognostic differentiation margins pyrimidines alone and plus/minus oxaliplatin, seem to be
between high, intermediate and low scores, their use might more limited with a higher likelihood for toxicity in older
be considered in complementing clinicopathological infor- patients.
mation on intermediate-risk stage II scenarios: i.e. to treat ! MSI/MMR status is the only validated molecular marker
T3 N0 classified as high risk by the signature, or for avoiding used in adjuvant decision making and should be deter-
chemotherapy in T4 N0 classified as low risk by the signa- mined in stage II CRC. In stage III, usage of MMR status
ture [II, C]. is limited to detect and identify Lynch syndrome [IV, A].
Immunoscore has been recently validated in a large ! DPD genotyping or phenotyping is strongly recommen-
prospective cohort of >2500 patients TNM stage IeIII.70 ded before initiating fluoropyrimidine-based adjuvant
Immunoscore was a strong predictor for time to recur- therapy according to regulatory bodies [III, A].
rence, OS and disease-free survival (DFS) (all P < 0.0001), ! Gene expression signatures are not recommended for
independently of patient age, sex, MSI and other existing routine practice due to lack of predictive value for
prognostic factors. Immunoscore had the highest relative chemotherapy benefit; however, clinicians and patients
contribution to the risk of all clinical parameters, including may consider their use to complement clinicopatholog-
the UICC TNM classification system.70 Therefore, Immuno- ical information in intermediate-risk stage II scenarios
score could help refine the prognosis of early colon cancer although their role in predicting chemotherapy benefit
patients in conjunction with the TNM scoring [III, C]. is uncertain [II, C].
However, its role in predicting chemotherapy benefit is ! Immunoscore could be considered to refine the prog-
uncertain and firm evidence of its prognostic role in a stage- nosis of early colon cancer patients used in conjunction
II-only dataset is currently lacking. with the TNM scoring and thus adjust the chemotherapy
Finally, ctDNA monitoring, also known as liquid biopsy, is a decision-making process in stage II and even in low-risk
promising tool under investigation to identify patients with stage III patients [III, C], although its role in predicting
high risk of recurrence after primary tumour resection. chemotherapy benefit is uncertain.
Indeed, ctDNA detection after stage II colon cancer resection
has been demonstrated to provide direct evidence of re-
sidual disease and to identify patients at very high risk of TREATMENT OPTIONS
recurrence.71 The results of ongoing trials investigating the
role of ctDNA as a tool to stratify patient’s risk of relapse and Stage III disease
to determine allocation to different adjuvant therapeutic The current standard of care for adjuvant therapy in stage III
strategies must be awaited before this is accepted in routine colon cancer is a combination of fluoropyrimidine and
practice. The CIRCULATE-IDEA and de Circulatie-Europa col- oxaliplatin. The benefit of these combinations over fluoro-
laborations seek to pool the data coming from the main pyrimidine monotherapy, the prior standard of care, has
national trials exploring ctDNA follow-up in the adjuvant been demonstrated in three landmark trials: MOSAIC,
setting. The results of this initiative will probably set the final NSABP C-07 and XELOXA. All showed significant improve-
role of ctDNA in the adjuvant decision-making process. ment in DFS compared with fluoropyrimidine as single
agent.72e74 The MOSAIC study used an infusional
Recommendations fluoropyrimidine regimen in both arms [leucovorin/5-
fluorouracil (LV5FU2) and leucovorin/5-fluorouracil/oxali-
! Adjuvant therapy options should be fully discussed with platin (FOLFOX)], the NSABP C-07 study used a bolus
the patient, taking into consideration tumour risk of fluoropyrimidine regimen in both arms [Roswell Park
recurrence, expected benefit from chemotherapy and and leucovorin/5-fluorouracil/irinotecan/oxaliplatin (FLOX)],
risk of complications. whereas the XELOXA study used a bolus fluoropyrimidine
! The risk of relapse after a colon cancer resection should regimen (Mayo Clinic or Roswell Park) compared with
be assessed by integrating the TNM staging, MMR/MSI capecitabine plus oxaliplatin (CAPOX). The MOSAIC and
status and number of lymph nodes sampled (#12) NSABP C-07 studies included both stage II and stage III
[III, A]. colon cancer, while the XELOXA study included only stage III
! Other additional clinicopathological features such as the colon cancer.
histological subtype and grading, lymphatic or venous or Although the chemotherapy regimens in the three
perineural invasion, lymphoid inflammatory response, studies were different, the addition of oxaliplatin resulted
involvement of resection margins and serum CEA should in a similar reduction in risk of recurrence in all three

Volume 31 - Issue 10 - 2020 [Link] 1297

Annals of Oncology G. Argilés et al.

studies (23% in MOSAIC and 20% in NSABP C-07 and patient has an ileostomy and in cases of renal insufficiency.
XELOXA). With longer follow-up, all three trials showed FOLFOX has higher reported neutropaenia rates. Immediate
improved OS from the addition of oxaliplatin with a risk oxaliplatin cessation following occurrence of grade >1
reduction of death of 16% in MOSAIC, 12% in NSABP C-07 neuropathy is recommended in all cases (whatever the
and 17% in XELOXA.65,72,75 However, a significant regimen and treatment duration) to avoid long-lasting
improvement in OS was only shown to be significant for symptomatic neurotoxicity that will impair the patient’s
stage III colon cancer. quality of life.
FOLFOX and CAPOX remain the current standard of care.
As the FLOX regimen results in increased incidence of Definition of risk groups in stage III. The IDEA study also
diarrhoea compared with FOLFOX or CAPOX, FLOX is not conducted an exploratory analysis based on risk sub-
currently recommended in clinical practice; in addition, groups. In the lower-risk subgroup (defined as patients
irinotecan, cetuximab and bevacizumab have not demon- with T1, T2 or T3 with N1 disease), 3 months of adjuvant
strated clinical activity in the localised setting and therefore therapy appeared to be sufficient, when CAPOX was
they should never be used as adjuvant treatment in this chosen [II, B]. In the higher-risk group (patients with T4 or
setting [I, E].76e80 N2 or both), 6 months of treatment may be necessary,
especially when FOLFOX is the chosen regimen, but also
IDEA collaboration, choice of regiment and treatment with CAPOX, which missed the non-inferiority margin on
duration of adjuvant treatment. The major cumulative this subgroup [II, B].
toxicity from a fluoropyrimidine/oxaliplatin doublet is sen- However, the panel believes that the establishment of
sory peripheral neuropathy. Worldwide, there have been six stage III risk subgroups should be used with caution, since
studies investigating whether 3 months of adjuvant this was a post-hoc analysis on the IDEA collaboration: T4
chemotherapy is non-inferior to 6 months of treatment, versus T1e3 and N2 versus N1 subgroups analyses were
with the aim of thereby diminishing the incidence of neu- pre-specified in the protocol but their combination in high-
ropathy and healthcare costs. These six trials have been versus low-risk subgroups was not, and moreover, its
examined prospectively by an international collaboration interaction test was not significant (P ¼ 0.11). Thus, the
and published as the IDEA study.81 In this pooled analysis, panel agrees that the established high- versus low-risk
12 834 patients with stage III colon cancer were randomised subgroups in stage III based on IDEA should have level of
to receive either 3 months or 6 months of a fluoro- evidence [V] (see Figure 3 for adjuvant treatment recom-
pyrimidine/oxaliplatin doublet (either FOLFOX or CAPOX); mendations in stage III).
the choice of regimen was mainly the clinician’s choice and
not randomised. The 3-year DFS rates were similar (overall: Stage II disease
74.6% and 75.5% for 3 months and 6 months, respectively) As already discussed, there are major and minor clinico-
but the pre-defined non-inferiority margin, accepting a 12% pathological factors that impact on the risk of relapse on
decrease as the upper limit of inferiority to be ruled out, stage II colon cancer. The presence of major factors
was not confirmed in the overall study population [HR 1.07; including pT4 stage or <12 lymph nodes assessed confers
95% confidence interval (CI) 1.00e1.15]. increased risk of recurrence, while the presence of other
However, sensory peripheral neuropathy grade 2 or additional risk factors is less significantly associated with
worse was significantly reduced from 34% with 6 months of risk of relapse.48e50 While follow-up is an option for low-risk
treatment to 11% with 3 months of treatment. stage II patients, chemotherapy is recommended for inter-
In the IDEA study, the treatment duration depends on the mediate- and high-risk patients [I, B].
choice of regimen. For patients receiving CAPOX, 3 months Although the de Gramont is the only regimen that has
treatment was non-inferior with 3-year DFS of 75.9% and demonstrated efficacy in the setting [I, B], capecitabine is an
74.8% for 3 and 6 months, respectively, whereas for FOL- option, especially with contraindications for insertion of a
FOX, 3 months treatment was inferior with 3-year DFS of central line [V]. It is also felt by the panel members that
73.6% and 76.0% for 3 and 6 months, respectively. There- patients with high risk, patients with pT4 and/or <12 lymph
fore, non-inferiority of the shorter regimen was seen for nodes or accumulation of several intermediate risk factors,
CAPOX (HR 0.95; 95% CI 0.85e1.06) but not for FOLFOX might be considered for the addition of oxaliplatin therapy
(HR 1.16; 95% CI 1.06e1.26). based on a trend to an increased benefit, although this did
Thus, both CAPOX for 3 months and FOLFOX for 6 months not achieve statistical significance in the stage II high-risk
can be recommended as adjuvant chemotherapy regimens subgroup analysis of the MOSAIC trial [I, B].72 For this
for stage III colon cancer [I, A]. It is important to mention high-risk population, the IDEA trial explored the optimal
that CAPOX and FOLFOX assignment in the IDEA trials duration of the oxaliplatin-based adjuvant treatment,
was not randomised, precluding any formal comparison finding identical results to those reported for stage III pa-
between the two regimens. tients, a non-proven non-inferiority for 3 months of treat-
CAPOX mitigates the need for central venous access and ment and there was a proven non-inferiority of CAPOX and
decreased neurotoxicity rates if 3 months is adequate but is inferiority of FOLFOX for 3 months when compared with
associated with more diarrhoea and hand-foot syndrome 6 months of FOLFOX,82 with all the limitations of these post
than FOLFOX; thus, it may be relatively contraindicated if a hoc analyses as stated before. The presence of MSI/MMR in

1298 [Link] Volume 31 - Issue 10 - 2020

G. Argilés et al. Annals of Oncology

Figure 3. Recommendations for adjuvant treatment of stage III colon cancer.

CAPOX, capecitabine plus oxaliplatin; FOLFOX, leucovorin/fluorouracil/oxaliplatin; MCBS, ESMO-Magnitude of Clinical Benefit Scale; MSI, microsatellite instability;
MSS, microsatellite stability.

localised disease confers better prognosis and less benefit delay of >8 weeks in starting adjuvant chemotherapy is
to adjuvant therapy so chemotherapy should be indicated associated with a higher relative risk of death (HR 1.20; 95%
with caution and always in combination with oxaliplatin51e55 CI 1.15e1.26, P ¼ 0.001).86 This observation has been
(see Figure 4 for integration of clinicopathological and confirmed by other groups.87,88 However, population-based
molecular factors with therapeutic recommendations). studies have shown that adjuvant chemotherapy might still
Lifestyle factors are likely to have an important impact on provide some benefit, even with delays up to 5e6
survival following adjuvant chemotherapy in either stage II months,89,90 but it seems that the benefit of adjuvant
or III patients, as reported for physical activity and nut chemotherapy is minimal or completely lost if treatment is
consumption.83,84 In addition, aspirin reduces the risk of started >6 months after surgery.
polyp formation and may also improve survival after adju-
vant chemotherapy in PI3K-mutated colon cancer patients
(approximately 20% of all patients).85 The ADD-ASPIRIN and Recommendations
ASPIK randomised studies are aiming to answer this ques-
! Combinations of fluoropyrimidines, either 5-FU or cape-
tion definitively.
citabine, and oxaliplatin constitute the bases for stage III
colon cancer adjuvant treatment [I, A; European Society
Timing of adjuvant chemotherapy for Medical Oncology-Magnitude of Clinical Benefit Scale
Delay between surgery and the beginning of adjuvant (ESMO-MCBS) v.1.1 score: B].
chemotherapy is a matter of debate. In view of the evi- ! The length of oxaliplatin-based adjuvant treatment of
dence, it is important to commence adjuvant chemotherapy stage III colon cancer based on the IDEA data may be
as soon as possible after surgery and ideally not later than 8 tailored to 3 or 6 months for CAPOX [I, A] or 6 months
weeks [II, B]. A meta-analysis of 14 studies showed that a for FOLFOX [I, A] also taking into consideration

Volume 31 - Issue 10 - 2020 [Link] 1299

Annals of Oncology G. Argilés et al.

Figure 4. Recommendations for adjuvant treatment of stage II colon cancer.

CAPOX, capecitabine plus oxaliplatin; CEA, carcinoembryonic antigen; FOLFOX, leucovorin/fluorouracil/oxaliplatin; MSI, microsatellite instability; MSS, microsatellite
stability.
a
For pT4 MSI: pT4 is a major risk factor but adjuvant chemotherapy benefit in the presence of MSI is uncertain.

pathological risk characteristics, patient comorbidity and ! For patients with low-risk stage II colon cancer, follow-up
risk assessment. is recommended [I, A].
! Further adaptation of the treatment according to risk ! For patients with intermediate risk (non-MMR/MSI þ any
subgroups: 3 months for CAPOX (T1e3 N1 disease), 6 risk factor except pT4 or <12 lymph nodes assessed), 6
months for CAPOX (T4 or N2 disease) or 6 months for months of fluoropyrimidines should be recommended
FOLFOX (T1e3 N1 or T4 or N2 disease) based on IDEA [I, B].
collaboration should be made with caution, since this ! Patients with high-risk stage II (pT4 or <12 lymph nodes
was based on a post-hoc analysis, non-significant for or multiple intermediate risk factors, regardless of MSI)
interaction [V]. may be considered for the addition of oxaliplatin [I, C].
! For patients not fit for or not tolerating oxaliplatin, either ! Patients with high-risk stage II colon cancer may be
capecitabine or LV5FU2 (de Gramont) infusion is accept- considered for 3 months of CAPOX, as the IDEA-pooled
able adjuvant regimens for a 6-month duration [I, A]. analysis showed non-inferiority of 3 months of CAPOX

1300 [Link] Volume 31 - Issue 10 - 2020

G. Argilés et al. Annals of Oncology

and inferiority of 3 months of FOLFOX when compared Regarding the timing and duration of follow-up, pro-
with 6 months of FOLFOX, with all the limitations of tocols need to be sensitive to the patterns of relapse of
post-hoc analyses [II, B]. colon cancer. Among recurring patients, 80% of relapses
! It is important to start adjuvant chemotherapy as soon as occur during the first 3 years and an additional 15% be-
possible after surgery and ideally not later than 8 weeks tween the 3rd and 5th year, which supports a more
[I, A]. intensive follow-up during the first 3 years and a stop after
5 years.91,96
FOLLOW-UP AND LONG-TERM IMPLICATIONS In addition to CEA testing and CT scans, colonoscopies
should also be included in the follow-up since metachro-
Follow-up nous primary cancer can be detected with an incidence of
Overall, between 30% and 50% of all patients treated for 0.7% within the first 2 years after curative surgery.97 How-
localised colon cancer will eventually relapse and die of the ever, there is no indication for intensive endoscopic follow-
disease.91,92 The main goal of follow-up protocols is detecting up. If a colon without tumour or adenoma is observed 1
relapse on an early basis, thereby maximising patient survival in year after resection, colonoscopy should be carried out af-
the metastatic setting. Systematic reviews have shown dispa- ter 3e5 years97 (see Figure 5 for colon cancer follow-up
rate results regarding the use of intensive follow-up as a tool to after curative resection).
increase OS.93,94 However, it has been shown that there is an
advance in the detection of recurrences [II, B] with intensive
follow-up.94 Detection of isolated local recurrences was Long-term implications/survivorship care plans
increased in the intensive group (15% compared with 9%, with CRC survivors represent the third largest group of long-term
risk ratio 1.61 and P ¼ 0.011), along with a small, non- cancer survivors in Western countries, w11% of this pop-
significant increase in the detection of hepatic metastases.94 ulation. For this group, additional post-therapeutic follow-
However, heterogeneity of the studies included in these up interventions have demonstrated to improve patient
meta-analyses does not allow precise assessment of algorithms outcomes.98 In this setting, the primary practitioner should
for optimal surveillance in clinical practice. Only trials including have a significant role in collaborating with the oncological
clinical assessment, CEA testing and/or liver imaging achieved teams.99,100
significant improvements in survival, though all studies Major elements in survivorship care are as follows:
considering liver imaging also included blood CEA monitoring.95
On one hand, CT scan including optimal liver assessment ! Prevention of recurrent and new cancer (classic end
has been shown to be more sensitive than ultrasonography point of follow-up).
(0.67 compared with 0.43) for liver relapse follow-up and, in ! Intervention for cancer sequelae and their treatment
addition, can detect chest recurrences. On the other hand, (rehabilitation).
liver MRI may be an alternative when a CT scan has shown ! Assessment of medical and psychological late effects
confusing liver lesions.96 (modern end point of follow-up).

Figure 5. Recommendations for follow-up after curative resection.

CEA, carcinoembryonic antigen; CT, computed tomography; mCRC, metastatic colorectal cancer.

Volume 31 - Issue 10 - 2020 [Link] 1301

Annals of Oncology G. Argilés et al.

! Health promotion (lifestyle promotion, comorbidity used to calculate scores for new therapies/indications
prevention, etc.). approved by the EMA since 1 January 2016 ([Link]
[Link]/Guidelines/ESMO-MCBS). The scores have been
Most long-term survivors of CRC report good quality of calculated by the ESMO-MCBS Working Group and
life following treatment, but several problems are still validated by the ESMO Guidelines Committee. Levels of
observed.101 A significant proportion of patients have evidence and grades of recommendation have been applied
persistent bowel dysfunction. It is important to refer for using the system shown in supplementary Table S3, avail-
dietary counselling and suggest use of over-the-counter able at [Link]
medications (e.g. fibre laxatives, stool softeners, antidiar- Statements without grading were considered justified
rheals). Colostomies and ileostomies represent also a standard clinical practice by the experts and the ESMO
source of physiological distress and disturbances at the level Faculty. This manuscript has been subjected to an anony-
of social functioning. Patients should be encouraged to take mous peer review process.
part in ostomy management programmes and psychological
distress management programmes must be recommended ACKNOWLEDGEMENTS
in case of discomfort with their body changes. The ESMO Guidelines Committee would like to thank the
Colon cancer survivors experience higher rates of sexual ESMO Faculty and other experts who provided critical re-
distress and psychological depression.101 Assessment of views of these ESMO Clinical Practice Guidelines.
distress should be considered, but evidence on the effec-
tiveness of psychosocial interventions among survivors of FUNDING
CRC is limited. Patients should be encouraged to maintain a
healthy lifestyle including exercise, quitting smoking, No external funding has been received for the preparation
avoidance of excessive alcohol intake and adoption of a of these guidelines. Production costs have been covered by
healthy diet rich in vegetables, fruit and berries adapted to ESMO from central funds (no grant number).
the remaining gastrointestinal function.102
DISCLOSURE
Recommendations GA has reported that he has acted as a consultant or
speaker for Amgen, Roche, Merck, Sanofi, Servier, Merck
Sharp & Dohme and Bayer. JTab has reported consultancy
! Intensive follow-up allows earlier detection of relapses in or speaker roles for Amgen, Roche, Merck, Celgene, Lilly,
patients at risk [II, B]. Sanofi, Sirtex, Pierre Fabre, Merck Sharp & Dohme and
! History and physical examination and CEA level determi- Servier. TI has reported honoraria for advisory boards from
nation are advised every 3e6 months for 3 years and Amgen, Bayer, Bristol-Myers Squibb, Celgene, Pierre-Fabre,
every 6e12 months at years 4 and 5 after surgery [II, B]. Roche and Servier. DH has received research funding from
! Colonoscopy must be carried out at year 1 and every Merck Serono. JTai has reported personal financial interest
3e5 years thereafter, looking for metachronous ade- from scientific consultancy roles for Array Biopharma,
nomas and cancers [III, B]. AstraZeneca, Bayer, BeiGene, Boehringer Ingelheim, Chu-
! CT scan of chest and abdomen every 6e12 months for the gai, Genentech, Inc., Genmab A/S, Halozyme, Imugene
first 3 years can be considered in patients who are at higher Limited, Inflection Biosciences Limited, Ipsen, Kura
risk of recurrence according to the TNM classification [II, B]. Oncology, Lilly, Merck Sharp & Dohme, Menarini, Merck
! Other laboratory and radiological examinations are of Serono, Merrimack, Merus, Molecular Partners, Novartis,
unproven benefit and must be restricted to patients Peptomyc, Pfizer, Pharmacyclics, ProteoDesign SL, Rafael
with suspicious symptoms [V, C]. Pharmaceuticals, F. Hoffmann-La Roche Ltd, Sanofi, Sea-
! Long-term follow-up, rehabilitation and survivorship care Gen, Seattle Genetics, Servier, Symphogen, Taiho, VCN
programmes should be implemented, aiming at detec- Biosciences, Biocartis, Foundation Medicine, HalioDX SAS
tion of recurrent or new cancers, assessment and man- and Roche Diagnostics and institutional financial support
agement of late and psychosocial effects and for clinical trials or contracted research for Agendia BV,
implementation of health promotion measures [III, A]. Amgen SA, Debiopharm International SA, Janssen-Cilag SA,
Mologen AG, Novartis Farmaceutica SA, PharmaMar, Roche
Farma SA, Laboratorios Servier SL and Symphogen A/S. RL
METHODOLOGY has reported leadership role for GISCAD. RS has reported
These Clinical Practice Guidelines were developed in advisory board/speaker for Pfizer, Novartis, Amgen, Merck
accordance with the ESMO standard operating procedures and Merck Sharpe & Dohme, advisory board for VCN-
for Clinical Practice Guidelines development ([Link] Biosciences, Agendia, Guardant Health, Roche Di-
[Link]/Guidelines/ESMO-Guidelines-Methodology). The agnostics, Ferrer, Ipsen, Roche Pharma and Lilly, speaker
relevant literature has been selected by the expert authors. role for AstraZeneca and Celgene, leadership role for and
An ESMO-MCBS table with ESMO-MCBS scores is included stocks from Sace Medhealth, grants/research support from
in supplementary Table S2, available at [Link] Roche Diagnostics, Novartis Farmaceutica, VCN Bio-
1016/[Link].2020.06.022.103 ESMO-MCBS v.1.1 was sciences, Merck KGaA, Roche Farma and Mologen. PLP has

1302 [Link] Volume 31 - Issue 10 - 2020