Clinical practice guidelines for the surgical management of colon cancer: a consensus statement of the Hellenic and Cypriot Colorectal Cancer Study Group by the HeSMO*

Background

Optimum therapeutic strategy and adequately executed surgery is best produced in volume-based referral centers by an adequately trained multidisciplinary team (MDT) which should include surgeons, radiologists, medical oncologists and pathologists. Further to centralization and adherence to clinical guidelines, oncological outcomes are expected to improve by national audit registries [2,4,6]. It should be mentioned that guidelines do not always derive from high quality level I data, and therefore should be applied with caution.

Clinical examination

Background

Physical examination and medical and family history of colorectal cancer, polyps and other cancers should be obtained. Further to assessment of the primary colonic tumor, total colonoscopy is mandatory to detect any synchronous lesions. Additional investigations like virtual colonoscopy or CT colonography could be helpful, even though they are not yet standard procedures. These could be valuable to precisely locate the tumor, particularly useful for the surgical approach, especially in patients who are candidates for a laparoscopic resection. They could also help detect other synchronous colonic lesions or polyps if colonoscopy could not explore the whole colon due to an obstructive tumor [7].

Image staging (Table 3)

Table 3 Initial staging of colon cancer

Background

Treatment strategy for colon cancer is guided by adequate staging. Complete colonoscopy and multi-detector computed tomography (MDCT) scan of the chest, abdomen, and pelvis should be performed. MDCT remains the main imaging modality for preoperative planning, metastatic liver lesion detection and tumor surveillance. MDCT accuracy rate for assessing lower stage lesions is not as good as that for advanced lesions. This discrepancy relates to the limited ability of MDCT to determine depth of bowel wall penetration. However, abdominal/pelvic MDCT has a high negative predictive value. The specificity for detecting lymph nodes involved with tumor is approximately 50%. In addition, the modality offers the ability of a rapid global evaluation and demonstration of complications (perforation, obstruction, etc.) that may not be clinically apparent [8,9]. Among patients with potentially resectable liver metastases and a negative initial chest x-ray, additional imaging with a chest CT may detect pulmonary metastases in up to 5% of patients [10].

Magnetic resonance imaging (MRI) has equal accuracy to MDCT for local staging of colonic neoplasms. Accuracy in identification of lymph node metastases is also equal to MDCT, and slightly superior for detection of liver metastases. MRI may be beneficial in determining involvement of the adjacent organs. MRI may also be considered in preoperative evaluation of patients with sensitivity to iodinated contrast material, particularly in the evaluation of the liver [8,11]. MRI and contrast-enhanced ultrasonography (US) should be considered as problem solving techniques for characterization of indeterminate liver lesions [8,9].

Computed tomographic colonography (CTC) can accurately identify all colorectal masses but may overcall stool as masses in poorly distended or poorly prepared colons. CTC has an overall staging accuracy of 81% for colorectal cancer and is superior to barium enema in visualizing colonic segments proximal to obstructing colorectal lesions. Furthermore, the method can identify synchronous lesions in patients with colorectal masses, and image the proximal colon in patients with obstructing colorectal lesions [8,12].

FDG-PET is not recommended for initial staging. It could be used in patients at high surgical risk when there is a strong probability of metastatic disease invisible on CT or MRI. However, the role of FDG PET/CT is not yet clear owing to the small number of studies [13]. Also, brain and bone scintigraphic scans are only indicated in patients with relevant symptomatology.

Preoperative laboratory assessment

Background

A complete blood count is necessary to determine levels of hemoglobin. If hemoglobin is less than 8 g/100 mL increased postoperative morbidity is expected. A hemoglobin level of above 10 g/100 mL prior to surgery is desirable. This can be achieved either with blood transfusion at least 2-3 days prior to surgery or with the administration of erythropoietin and iron intravenous infusion at least two weeks preoperatively. There is evidence that blood transfusion impairs immune response of the patient to malignant process of the disease, which may translate to worse long-term oncological outcomes [14,15]. For this reason, erythropoietin and iron infusion is recommended as the safest choice in oncological terms, although relative evidence is not sound and recommendation level is low [16]. Defective clotting mechanisms should be corrected accordingly before surgery.

A baseline determination of serum carcinoembryonic antigen (CEA) level is necessary for the indirect detection of completeness of surgery, namely increased preoperative CEA levels should be normalized after curative surgery. Also according to a recent review [17], postoperative increase of serum CEA is a highly specific but insufficiently sensitive factor for the detection of local or distant recurrent colorectal cancer. A cut-off value of 2.2 ng/mL may provide the ideal balance of sensitivity and specificity. Therefore, serial serum CEA determination is highly recommended as a first-line surveillance test.

Bowel preparation

Background

There is substantial evidence that mechanical bowel preparation prior to elective colectomy does not offer any advantage over surgery without bowel preparation [18-20]. As a matter of fact and according to a recent meta-analysis, mechanical bowel preparation may be associated with increased rate of anastomotic leak and wound infection [18].

Enhanced recovery programs

Background

Implementation of enhanced recovery programs, so called “fast-track”, in colorectal surgery for both benign and malignant diseases reduces physiological and psychological stress, accelerates normalization of gastrointestinal function, improves postoperative physical status, and most importantly is associated with less morbidity and shorter postoperative hospital stay. In addition, implementation of those programs does not necessitate specific equipment nor does it require increased costs. Therefore, implementation of “fast-track” is strongly recommended in units with motivated and adequately trained personnel [21-23].

Surgical treatment

Background

Malignant polyp

Treatment of the malignant colonic polyp depends on clinical and histopathological features:

An intense schedule of follow up is recommended in patients with a removed malignant polyp and no further treatment (Table 4).

Table 4 Management of the malignant colonic polyp

Resectable non-obstructing lesion (Fig. 1)

Figure 1 Strategy and treatment algorithm of non-metastatic colon cancer

CT, computed tomography; CTx, chemotherapy

Background

For early cancer stage 0 or partly stage I (T1) local excision by means of the colonoscope could be considered, particularly in patients with increased co-morbidities. If histology shows clear margins of resection, well-differentiated tumor (G1, G2) and no lymphatic invasion, an expectant policy is recommended, as local recurrence is not very likely and lymph node metastasis may occur in only up to 4%. In case histology shows incomplete resection margins, a poorly differentiated lesion (G3, G4) or lymphatic invasion, surgical curative resection should follow, as local recurrence is very likely and lymph node metastasis may occur in up to 20% of the cases [24].

For resectable colonic carcinoma, the oncologically optimal surgical procedure is a curative (R0) colectomy with adequate proximal and distal resection bowel margins, and en-bloc complete removal of the respective to the resected segment mesocolon (Complete Mesocolic Excision - CME) with all regional lymph nodes [25-27]. Based on the fact that potential involvement of pericolic lymph nodes does not extend the 8 cm proximal and distal to the tumor baring bowel segment, bowel resection margins should be at least 10 cm, unless this is restricted by the exact location of the tumor or/and type of colectomy [28,29]. As a general rule, proximal ligation and division of the vascular stems supplying the specimen to be resected (central vascular ligation, CVL) ensures CME and the highest possible retrieved number of lymph nodes [26,27].

For tumors situated at the cecum and ascending colon, CVL involves the ileocolic vessels and the right branches of the middle colic vessels. For tumors situated at the right side of the transverse colon, CVL involves the ileocolic and the middle colic vessels. For tumors situated at the middle and left transverse colon and the upper descending colon, CVL involves the middle colic vessels, the ascending branches or the stalk of the ileocolic vessels and the ascending branches of the left colic vessels. For tumors situated at any site from the descending colon to the rectosigmoid junction, CVL involves the division of the inferior mesenteric artery at 1 cm distal to its origin from the aorta and the inferior mesenteric vein just below the lower border of the pancreas [26,27].

Resection should be complete in terms of removal of all regional to the resected bowel segment lymph nodes. A resection is considered incomplete (R2) if involved lymph nodes are not removed. The number of the lymph nodes removed depends on the location of the tumor [30]. In general, right colon segments tend to contain much higher numbers than the left ones. According to UICC recommendations [31], at least 12 lymph nodes are required to be examined for the staging of the disease. If all lymph nodes examined are negative but <12 in number, staging is not optimal and safe. The accuracy of staging of colorectal cancer parallels the number of removed lymph nodes [32]. There are two additional reasons emphasizing the significance of the number of removed lymph nodes: i) increased number of removed lymph nodes is associated with improved survival, irrespective of the nodal status [33]; ii) the ratio of the metastatic to total number of removed nodes is inversely related to recurrence and overall survival [34]; and iii) increased absolute number of negative retrieved nodes is associated with better oncological outcomes even in stage III disease. The latter two stand true only when the number of examined nodes is >12 [30,35-39].

Anastomosis to reestablished bowel continuity could be performed with use of sutures or stapling devices. Current evidence shows that there are no differences in the anastomotic leak or stenosis rates between the two approaches [40]. Possibly stapled ileo-colic anastomosis may be associated with fewer anastomotic leaks than those hand sewn [41]. There is also no difference in the anastomotic complication rate between single and double-layer sutured anastomosis [42].

Laparoscopic approach

Background

Laparoscopic colectomy is nowadays the alternative to the open approach. There is substantial evidence based on several comparative studies that the laparoscopic approach for colon cancer surgery is as effective and as safe as the open one. Several meta-analyses and systematic reviews [43-49] including among others three large multicenter randomized comparative trials: the COST trial in the USA involving 872 patients [50-52], the CLASICC trial in the UK involving 794 patients [53,54], and the COLOR trial in Northern Europe involving 1248 patients with colon cancer [55], plus several other single-center comparative randomized trials [56-59] clearly show that the laparoscopic approach is associated with faster recovery, less postoperative pain and use of narcotic analgesics and less immediate postoperative morbidity as compared to the open approach. Furthermore, quality of surgery, as depicted in the percentage of involved resected margins and the number of retrieved lymph nodes, is similar between the two approaches. Also, local recurrence and overall and disease-related survival are similar between the two approaches. Finally, and in the long run, there is less readmission rate due to obstructive ileus [60] and lower incidence of incisional hernias [61,62] after the laparoscopic than after the open approach. It has been recently shown that laparoscopic colectomy works perfectly with enhanced recovery programs, offering even better faster recovery and less immediate postoperative morbidity [63-65].

It is recommended that laparoscopic colectomy for cancer should be performed by experienced surgical teams with adequate case volume, a necessary factor for improved outcomes [66]. Conversion to open, particularly when performed at a late stage of the procedure, may be associated with increased postoperative morbidity as compared to both the laparoscopically completed and the open approach. The main predictive factors for conversion are the T4 large tumors, the increased ASA condition of the patient, obesity and surgeon’s limited experience [67]. Preoperative image staging is mandatory to identify T4 tumor or tumors >8 cm which should be amenable only to the open approach. Previous surgery and advanced age are not contraindications to laparoscopy. Also, it should be considered that obesity, though not an absolute contraindication, is associated with increased rate of conversion. Tumors not invading the bowel wall and unlikely to be visually identified at laparoscopy should be marked with Indian ink prior to surgery or localized with an on -table colonoscopy. The laparoscopic approach is not recommended for carcinomas located at the transverse colon, because dissection of the middle colic vessels is very difficult and laborious and quality of specimen is not optimal [68]. The laparoscopic approach is not recommended in acutely perforated or obstructing colonic tumors.

Macroscopic specimen assessment

Background

Quality of the resected colonic specimen by either approach should be assessed macroscopically. The procedure involves spread but not stretching of the specimen on paper or towel and at first photographed at both sides. Then completeness of mesocolon resection is graded (mesocolic, intramesocolic and intramuscularis). A mesocolic resection is associated with better survival as compared to intramuscularis [25-27]. Also, with the use of a ruler: i) the whole bowel and the proximal and distal to the tumor length, as well as the tumor length; ii) the distance of the central vascular tie to tumor; and iii) the shortest distance from the central vascular tie to the bowel wall are measured [27].

Resectable non-obstructing synchronous bowel carcinomas

Background

Synchronous carcinomas of the colon are resected in the form of one specimen and one anastomosis is established. Segmental colonic resection with multiple anastomoses should be avoided, as they are associated with increased morbidity. When one of the carcinomas is located at the right colon, an extended right hemicolectomy is recommended. Likewise, when two lesions are located at the left colon, a left hemicolectomy is recommended. The procedure can be performed by either open or laparoscopic approach in selected cases.

Obstructing carcinoma

Background

For resectable colon cancer causing acute complete obstruction, resection with one or two stage anastomosis is recommended [69-72]. If the condition of the patient does not permit surgery, stenting, as a bridge procedure, for the alleviation of the obstruction first and curative resection after two weeks is the management of choice [73-75]. However, evidence for the superiority of stenting in malignant colonic obstruction over emergency surgery is not sound, apart from a shorter hospital stay, a less procedure duration and less blood loss [76].

If the obstructing carcinoma is unresectable, palliative surgery in the form of defunctioning proximal stoma or insertion of a stent is recommended. If the general condition of the patient permits it, chemotherapy and/or radiotherapy could be added [77].

Hereditary non-polyposis colon cancer (HNPCC)

Background

Considering that HNPCC patients are younger in age than those with sporadic colon cancer and probability to develop metachronous lesions is very high, a total colectomy is recommended [78-80].

Logoregional recurrence

Background

Definition, incidence, risk factors

Reports on local recurrence of colon cancer after curative resection are limited and, as a result, incidence, risk factors and clinical presentation of the condition are not well defined. Even definition of locoregional recurrence is poor, and may involve primarily the anastomosis, but also the mesenteric lymph nodes, the peritoneum or even the retroperitoneal space. A more comprehensive definition of locoregional recurrence is the recurrence at the abdominal quadrant of primary tumor location [81-84].

According to a population-based study [83], the incidence of recurrence after potentially curative resection of colon cancer was 11.5% at 5 years. Most of the recurrences were anastomotic. According to the same study, emergency conditions, namely perforation or obstruction, at the initial presentation, are independent risk factors for recurrence, possibly because of technical difficulties and suboptimal primary surgery, as most of these procedures are performed by non-colorectal surgeons [83,85-88]. Also, advanced T and N stage, location of the primary tumor at the hepatic flexure or sigmoid, poor differentiation of the primary tumor, and suboptimal primary surgery with low number of retrieved lymph nodes are considered risk factors of recurrence [81-84]. More than 80% of the cases are discovered by symptomatology. In the remaining cases recurrence is found at the regular follow up, as anastomotic recurrence at endoscopy and increased serum CEA [81-84].

Imaging

CT is not very accurate for early detection of colon cancer local recurrence, due to the distorted local anatomy after operation. The reported sensitivity reaches only 53% for CT, while FDG-PET shows a much better sensitivity of 93% [83]. The typical PET/CT appearance is of a hypermetabolic soft tissue mass or subtle wall thickening at the anastomotic site, often identified by a surgical ring of radio-opaque staples. Although colonoscopy would be the ideal technique for diagnosing and confirming anastomotic site recurrences, FDG PET/CT imaging can prove to be an excellent noninvasive modality when such recurrences are suspected, specifically in the cases of extramural and/or nodal infiltration. At present, whole-body (18)F-FDG PET/CT is an advanced diagnostic imaging technique in detecting locoregional recurrence and metastases in postoperative patients with colonic carcinoma for its higher sensitivity and specificity [89-91]. MRI does not seem to offer an added value, since its sensitivity in identifying early postoperative colon cancer recurrence is relatively low. Whole-body diffusion-weighted MRI is being explored for recurrent colon cancer, however, there is currently no evidence to suggest that this functional technique can replace PET/CT [83,89].

Management

A MDT should evaluate patients with recurrence, although this is not usually the case, according to Sjoevall et al [83]. Surgery is offered in almost 60% of the cases, in combination with radiotherapy and chemotherapy. Palliative measures, including chemotherapy and radiotherapy, are reserved for the remaining cases [81-84].

Potentially curative resection can be achieved to approximately one fifth of those amenable to surgery. Most of potentially curative resections are elective. Only complete resections offer a survival rate of approximately 40% at five years, whilst none of the patients with recurrent colon cancer and incomplete or no resection survive 5 years [81-84].

Pathology

Background

If not ex vivo in the theater, macroscopic assessment of the fresh specimen is processed in the laboratory as described before. The lateral resection margin of the fresh surgical specimen must be inked. The surgical specimen is opened leaving intact the tumor area and 2 cm below and above it and is then fixed in formalin solution for 48 h.

The histology report must include:

A. Gross description which involves length of surgical specimen, tumor size (3 dimensions) distance from proximal or distal margin, depth of invasion, tumor perforation, other lesions not related with the tumor (Crohn’s disease, ulcerative colitis, adenomatous polyp, familial adenomatous polyposis), and total number of lymph nodes. The distance of direct tumor spread outside the muscularis propria should be recorded and the area in which tumor spreads closest to the lateral resection margin should also be identified macroscopically.

B. Microscopic description: Histologic type. The main histologic types in WHO classification are adenocarcinoma, mucinous adenocarcinoma (>50% mucinous), signet ring carcinoma (>50% signet ring), squamous carcinoma, adenosquamous, small cell, medullary and undifferentiated carcinoma. Although most histological types do not have any proven prognostic significance there are exceptions. Signet-ring and small cell carcinomas have poor prognosis. Mucinous and medullary carcinomas, when associated with microsatellite instability (MSI), have a favourable prognosis [92,93].

Histologic grade. Currently, a 2-tiered grading system is used (low and high grade). The system is based on the proportion of gland formation and in this way the inter-observer variation is avoided. Low grade has a proportion >50% glandular formation and in this grade the well and moderately differentiated carcinomas are included. In the high-grade category the poorly differentiated and undifferentiated carcinomas are included (<50% glandular formation) [92-94].

Lymph nodes. All lymph nodes found in the surgical specimen should be sampled. It has been shown that a minimum of 12 lymph nodes must be found to predict the real lymph node status. The interpretation of the discrete nodules of tumor in the adipose tissue on microscopic examination is many times problematic. According to the old guideline, extramural tumor nodules measured >3 mm in diameter but lacked evidence of residual lymph node tissue were considered as positive lymph nodes. According to the updated guideline, a discrete extramural invaded nodule with smooth contours irrespective of size is considered as positive lymph node [92-95]. Extramural and extranodal tumor deposits at the mesenteric fat are considered as remote metastatic disease and carry a poor prognosis [96].

Blood, lymphatic vessel invasion, perineural invasion. Several studies have shown that extramural vascular invasion (blood or lymphatic) is of strong prognostic significance and is associated with increased risk of liver metastasis. The prognostic importance of involvement of small vessels in the submucosa has also been well documented in the polypectomies for malignant polyps and is associated with risk of lymph node metastasis. Extramural vascular invasion is recorded when tumor is present within a space lined by endothelium and/or surrounded by muscle, and when inside the space erythrocytes are observed. Studies have shown that the detection of venous invasion depends on the number of blocks taken from the tumor periphery. The College of American Pathologists recommends 3-5 blocks from the deepest part of the tumor to be examined [92,93]. Also, it has been shown that perineural invasion is an independent indicator of poor prognosis [92,93].

Tumor infiltrating lymphocytes. The intratumoral lymphocytic infiltration is associated with MSI, medullary architecture and is considered as a favorable prognostic factor [92,93].

pTNM classification. Colorectal cancer is classified according to the pTNM system [92-94] (Table 5).

Table 5 pTNM classification

Residual tumor classification. Surgical margin status should be reported. For the resection margins after surgery, the R classification system is advocated. Four grades are in use: Rx (presence of residual tumor cannot be assessed); R0 (no residual tumor - distance from closest involved margin must be reported; R1 (microscopic residual tumor); and R2 (macroscopic residual tumor) [92-94].

Assessment of pT1 colorectal tumor. pT1 tumors invade the muscularis mucosa and submucosa without invasion of muscularis propria. This group of tumors is often encountered in early adenocarcinomas developed in adenomatous polyps. Histopathology report must include: histological grade, distance of tumor from the resection margin, vascular or lymphatic invasion and depth of invasion into submucosa. According to Kikucki levels, the invasion of submucosa is graded in three levels: sm1 (superficial part of submucosa), sm2 (middle part), and sm3 (deep part). Adenocarcinoma spreading to within 1mm or less of the surgical or endoscopic resection, presence of lymphatic or vascular invasion and high-grade differentiation, as well as mid and deep third invasion of the submucosa are findings suggesting an increased risk for presence of lymph node metastasis [92,93,97].

Follow up

Background

Patient follow up depends on stage of the disease, quality of surgery, and amenability for intervention with either resection of recurrent disease or consideration of further systemic therapy. Four recent meta-analyses have shown that intensive follow up improves overall survival by 7-13% and is now considered as standard. Generally, valid assessments are 3-monthly clinical visits for the first three years, followed by 6-monthly visits for further two years with clinical examination, evaluation of long-term toxicities (polyneuropathy after oxaliplatin), and CEA testing [4,98-100].

Complete colonoscopy must be performed at initial diagnosis, after three and afterwards every five years. If the diagnostic colonoscopy was incomplete, this should be repeated at 6 months after surgery [4,99,100]. If follow-up colonoscopy shows advanced adenoma it should be repeated in one year.

In patients with high-risk disease, CT scan of the chest and abdomen every 6-12 months could be considered, although such close follow up should be confined to patients possibly amenable to resection of hepatic or pulmonary recurrence. Follow-up CT scans should be performed with the same imaging protocols and contrast phases of enhancement [4,99]. If MRI was used for the initial staging, MRI should also be used for the follow up, because CT images cannot be compared to MRI images due to different sensitivity/specificity [4]. Liver contrast-enhanced US could substitute for abdominal CT scan regarding follow up of liver metastases, particularly in young patients with no evidence of extrahepatic disease [4,101]. Finally, routine PET-CT scanning is not recommended for surveillance [4,99].