Colorectal cancer: what the clinician wants to know.

Colorectal cancer is a common and lethal disease. The adenoma-carcinoma sequence offers a window of opportunity in which the precursor lesion or early carcinoma can be removed endoscopically to prevent systemic disease. New and advanced techniques to improve endoscopic detection of precursor lesions are being developed. Other, less invasive screening methods are currently being developed and may become of use for population-based screening in the near future. Recently, important developments in the treatment (both surgical and chemotherapeutic) of colorectal cancer have occurred. The extent of the disease (stage) forms the basis for therapeutic decisions and accurate imaging is crucial. International Cancer Imaging Society.

Introduction

Colorectal cancer is a common and lethal disease. Every year, more than 945000 people develop colorectal cancer worldwide, and around 492000 patients die [1]. Despite these sobering statistics, many advances have been made in different aspects of this disease in recent years. In this review we focus mainly on the diagnostic and therapeutic aspects of colorectal cancer.

Pathogenesis of colorectal cancer

The majority of cases of colorectal cancer develop sporadically (88%–94%). The remainder occur in high-risk groups like hereditary cancer syndromes (e.g. hereditary non-polyposis colorectal cancer (HNPCC) and familial adenomatous polyposis) and long-standing inflammatory bowel disease. Furthermore, patients with previous adenomatous polyps or cancer of the colon are also at increased risk for developing colorectal cancer.

Most colorectal carcinomas are thought to arise in pre-existing adenomatous precursors (polyps) that form in the colon when normal mechanisms regulating epithelial renewal are disrupted: the adenoma-carcinoma sequence. As the lesion progresses, genetic changes in various tumour suppressor genes and oncogenes accumulate [2]. However, only a small proportion of all adenomatous polyps progress to cancer and furthermore, the speed of progression varies. For example, the adenoma-carcinoma sequence for sporadic cancers takes an estimated 10–15 years, whereas adenomas in patients with HNPCC might proceed to cancer at a significantly higher speed. To further complicate matters, there is increasing evidence of alternative pathways leading to colorectal cancer and the serrated pathway might serve as an explanation for the occurrence of colorectal cancer in hyperplastic polyps [3]. Thus, it is becoming clear that colorectal cancer is not a single, but rather a very heterogeneous disease from both a genetic and a clinical point of view.

Window of opportunity

The adenoma-carcinoma sequence offers a window of opportunity in which the precursor lesion or early carcinoma can be removed endoscopically to prevent systemic disease. Indeed, there is a significant reduction in the incidence of colon cancer in screened populations by white light colonoscopy plus polypectomy [4].

However, the question arises which polyps should be identified and removed? The National Polyp Study Workgroup introduced the concept of adenomas with advanced pathologic features that serve as risk factors for developing future colorectal cancer [4]. These factors are: adenomatous polyp >1 cm in diameter, polyp with high-grade dysplasia or invasive cancer. Furthermore, villous histology and increasing polyp size also correlate with the development of colorectal cancer [5]. The risk for development of colorectal cancer after removal of small tubular adenomas seems to be low [5, 6]. However, in clinical practice most endoscopists also remove smaller adenomas.

For decades hyperplastic polyps were thought to cause no harm. However, now that the serrated pathway for the development of colorectal carcinoma has been described, there is no clear consensus if these polyps should also be removed. The hyperplastic polyps that are large in size and occur in the proximal colon seem to carry a higher risk and should be considered for removal [3].

Diagnosis of colorectal polyps and cancer

Surveillance and screening

Colonoscopy is traditionally considered the optimal examination for the detection of polyps and colorectal cancer. During colonoscopy lesions can be localized throughout the colon, biopsies can be taken for histologic diagnosis, synchronous lesions can be detected and adenomatous polyps and early cancers can be removed in their entirety. However, colonoscopy is not infallible and both polyps and cancers can be missed [7]. The miss rate is inversely related to the size of the lesion and influenced by both the endoscopist and patient-related factors. Furthermore, some cancers also apparently arise from adenomas that are not polypoid. Flat and depressed lesions may account for 12%–40% of all adenomas and early colorectal carcinomas [8, 9]. These lesions have a surprisingly high rate of submucosal invasion even when they are small and, needless to say, are more difficult to detect by colonoscopy. Patients with long-standing ulcerative colitis or Crohn’s disease have an increased risk for flat dysplasia and cancer that may develop within fields of transformed mucosa which is usually difficult to detect endoscopically [10]. Furthermore, it would be very attractive to be able to immediately differentiate between neoplastic and non-neoplastic lesions during colonoscopy. In this way, only neoplastic lesions could be removed during the same colonoscopy session and other lesions might be left in situ, thus reducing time, costs and risk of the procedure.

Chromoendoscopy (application of dye during endoscopy) in combination with magnification endoscopy is one of the techniques with proven success. Rembacken et al. and many others have reported this technique to be helpful for both detection and detailed morphological assessment of colorectal lesions [9]. Kudo et al. performed pioneering studies, combining high-magnification colonoscopy with chromoendoscopy for the in vivo prediction of histology using a pit-pattern classification [11]. Pan-colonic chromoendoscopy showed a significant increase in diminutive adenoma detection [12, 13] and detection rates of dysplasia in patients with ulcerative colitis [14]. However, many more, novel endoscopic techniques are currently under investigation and offer promising means of improving endoscopic detection of precursor lesions and colorectal cancer. Amongst these are narrow band imaging, videoautofluorescence endoscopy and confocal microendoscopy. The most successful endoscopic method will probably be a combination of techniques that can both provide wide-area surveillance (a ‘red flag technique’ like fluorescence endoscopy) and, once a target-lesion is identified, detection methods like narrow band imaging or confocal microendoscopy that can be used to further identify the lesion [15].

Another attractive option may be the coupling of fluorescent dyes to tumour-related antigens as they are in every day use in most pathology labs. Immunoscopy is performed with the use of injected probes consisting of small molecules (enzyme substrates, receptor ligands, monoclonal antibodies, peptides) tagged to a fluorescent dye to achieve high-affinity binding to specific biochemical and molecular markers of disease [16]. In a German study, a fluorescein-labelled monoclonal antibody against CEA (carcinoembryonic antigen) was applied directly onto the mucosal surface during conventional colonoscopy [17]. Fluorescence in vivo was present in 19 out of 25 carcinomas and in 3 of 8 adenomas. However, the technique failed in the presence of mucosal ulceration or bleeding. The development of better optical tumour-specific imaging agents could have far-reaching potential in real-time polyp and cancer detection and will definitely complement existing fluorescence endoscopy but possibly also other imaging modalities.

The field of immunoscopy is very close to molecular imaging or may even be considered a form of molecular imaging. In the future other techniques such as magnetic resonance (MR) imaging may be more appropriate than colonoscopy and one could imagine an MR-colography which not only depicts possible polyps but also differentiates their nature through molecular imaging techniques. This would obviate the need for optical colonoscopy as a diagnostic method. Unfortunately, the road to these techniques still seems to be long.

However, in recent years a lot of research has been done on the role of virtual colonoscopy in diagnosing colorectal polyps or cancer. A meta-analysis on 33 studies of 6393 patients showed that computed tomography (CT) colography was highly specific, but its sensitivity was heterogeneous [18]. However, sensitivity improved with increasing polyp size. The results of MR-colography are less good than CT-colography, but this technique has no radiation exposure, which is of value in the choice of a screening modality. In virtual colonoscopy, patient acceptance seems better than for colonoscopy, and might be improved with less intensive bowel preparation. Important drawbacks of these techniques are the impossibility of obtaining histological biopsies and/or removal of the lesion and therefore this technique seems less attractive for surveillance of high-risk patients.

Population-based screening

Research has shown that population screening by testing for small, invisible traces of blood (faecal occult blood test, FOBT) can reduce mortality from colorectal cancer by 15%–33% [19-21]. Besides FOBT there are other, possibly more suitable methods for screening. Endoscopic screening is expected to reduce deaths to a greater extent than FOBT, but is an invasive investigation that is onerous for the patient and can lead to complications. Development of new, non-invasive imaging techniques like CT-colonography and MR-colonography, and non-invasive screening techniques such as detection of changes in DNA or proteins specific for cancer (proteomics) hold much promise for the future.

Staging

Once a cancer is diagnosed, the extent of the disease should be determined to form the basis for therapeutic decisions. In addition to physical examination, abdominal ultrasound and chest radiography are routinely performed to determine the extent of the disease. The necessity for routine preoperative CT scans is a matter of debate because this method alters the surgical approach in only a few cases [22].

In the case of rectal cancer preoperative knowledge of the depth of invasion and nodal status is critically important for the planning of treatment. Local staging can be done by endorectal ultrasound, CT or MRI. Positron emission tomography (PET) is valuable for detection of recurrent colorectal cancer, but seems to have only little effect on staging of primary cancer [23].

Detection of liver metastases can be performed by either CT or MRI. PET seems to hold promise but its use has not yet been proven in a randomized study [24].

Treatment

There are several modalities for the treatment of colorectal cancer and the decision for a certain therapy relies on the stage of the disease.

Endoscopic therapy

Since there are no lymphatics above the muscularis mucosae, many early malignant lesions can be managed endoscopically [25]. Thus, high-grade dysplasia (sometimes called carcinoma in situ) and intramucosal carcinoma are considered non-invasive if these lesions are found within a resected polyp; they require no further treatment if the resection margins are free of cancer. However, when histology of the cancer shows poor differentiation, lymphatic or vascular invasion, cancer at the resection margin, invasion of the submucosa, invasive carcinoma in a sessile polyp or incomplete polypectomy, there is a high risk of residual cancer and surgery should be considered.

Surgery for primary tumour

Traditionally, surgical resection of colonic cancer is based on resection of the intramural tumour together with wide excision of the area of regional lymph drainage, i.e. the mesentery. Hence, for tumours in the caecum, ascending colon or hepatic flexure, a right hemicolectomy is performed. Tumours in the transverse colon are resected through a transverse colectomy or alternatively, by extended right hemicolectomy. For tumours in the splenic flexure or descending colon, a left hemicolectomy is recommended, whereas for sigmoid tumours, a sigmoid colectomy is carried out. Localization of the tumour using double contrast barium enema, white light colonoscopy or virtual colonoscopy is therefore important to define the type and extent of surgical resection. In case of invasion of neighbouring structures or organs, such as ureter, bladder, small bowel or abdominal wall, the colon is resected en bloc with the adjacent structures in an attempt to achieve tumour negative margins. CT scan of the abdomen is helpful in identifying patients with such locally advanced tumours. It should be noted that nowadays, laparoscopic colectomy is equivalent to open colectomy in terms of oncological outcome [26].

Curative treatment of rectal cancer basically consists of local excision, low anterior resection (LAR) or abdominal perineal resection (APR). Full thickness local excision can be performed using the transanal endoscopic microsurgery (TEM) procedure, for superficially invasive cancers (T1). Although the recurrence rate after TEM is higher than after radical resection for more invasive tumours, TEM offers an alternative in patients with significant co-morbidity [27].

The prognosis of patients with rectal cancer has greatly improved by the concept of total mesorectal excision (TME) in combination with LAR or APR, which reduces local recurrences and perioperative morbidity [28]. A standard TME encompasses excision of the rectum including the complete rectal mesentery, proximal and distal to the tumour. The superior results of TME are attributed to improved lateral clearance and decreased risk of tumour spillage from a disrupted mesentery. Using this technique, local failure rates have dropped from 25% on average, to 4%–7% for Dukes B and C tumours [28].

APR entails rectal excision with complete proctectomy using both an abdominal and perineal approach and requires a permanent colostomy. APR is indicated in patients with rectal tumours involving the anal sphincter musculature or rectovaginal septum, or in patients with poor faecal continence. Otherwise, rectal excision with preservation of the sphincter (i.e. LAR) is the treatment of choice. Large invasive tumours of the distal rectum may be reduced by preoperative radiotherapy or neoadjuvant chemotherapy, converting a planned APR to a sphincter-sparing LAR. Hence, preoperative imaging is important to evaluate rectal tumours in terms of T staging and assessment of local, pelvic invasion.

Surgery for colorectal liver metastases

Partial liver resection remains the gold standard for curative treatment of patients with colorectal liver metastasis leading to a 5-year survival rate of 20%–40% [29]. The large majority (80%–85%) of patients with liver malignancies, however, are unresectable as a consequence of multifocal intrahepatic disease, extrahepatic disease, inadequate functional hepatic reserve, inability to obtain an optimal (0.5 cm) tumour free margin, or involvement of the confluence of the portal vein. It is the primary goal in the work-up of potential candidates for surgical resection, to reliably assess these criteria.

Selection of patients eligible for partial liver resection is nowadays based on the following four criteria: (1) Is the patient fit to undergo major upper abdominal surgery? (2) Are there no extrahepatic metastases (with the exception of lung metastases)? (3) Are all lesions resectable with a tumour-free margin >0.5 cm? (4) Is the volume of the remnant liver sufficient for adequate postoperative liver function?

Accurate imaging is crucial to answer the last three questions. CT and MRI are commonly used to assess the number and location of liver and lung lesions, and possible lymph node metastases. In the presence of lung metastases, pulmonary resection is carried out prior to abdominal exploration and liver resection. When planning liver resection, the segmental division of the liver as described by Couinaud is used to define the borders and extent of liver resection.

When the liver parenchyma is normal, which is usually the case in patients with colorectal liver metastases, up to 70% of total liver volume can be resected without increased risk for postoperative liver failure. The remnant liver has a unique potential to regenerate after resection, thus increasing in volume and function in the postoperative period. In specialized centres, the overall mortality of liver resections for liver metastases is 2%–5%. CT-volumetry of the (remnant) liver is therefore important to select candidates for liver resection. Alternatively, the volume of the future remnant liver can be calculated using patient characteristics. Although frequently used in the Far East for assessment of patients with hepatocellular carcinoma, the indocyanine green (ICG) clearance test is not routinely used in Europe for evaluation of liver function. In recent years, functional imaging of the liver has been examined using scintigraphic techniques (Tc-99m-galactosyl serum albumin or Tc-labelled Mebrofenin), providing simultaneous morphologic and physiologic information of liver segments [30].

Several strategies can be devised to resect liver metastases in patients initially considered unresectable. With neoadjuvant chemotherapy regimens, liver metastases can be downstaged thus offering survival benefits after resection comparable to patients resected on first presentation [31]. In the case of insufficient remnant liver volume, portal vein embolization of the affected liver lobe may be applied, resulting in ipsilateral atrophy and contralateral, compensatory hypertrophy of the future remnant liver [32, 33]. Alternatively, in order to preserve as much functional liver parenchyma as possible, liver resection can be combined with local ablative therapies such as radiofrequency ablation [34].

The “two staged” resection entails resection of the liver in two phases with a time interval to allow regeneration of the resected liver before the second resection is undertaken [35].

Chemotherapy

During the past 15 years, sequential advances in chemotherapy after surgical resection (adjuvant chemotherapy) have had a clear and substantial benefit, with the 4-year rate of overall survival approaching 80% [36].

The use of adjuvant fluorouracil-based chemotherapy in patients with Dukes stage III colon cancer is thought to be standard care, but is not routinely recommended in stage II colon cancer [37]. The prospects for a higher rate of cure in patients with locally advanced colorectal cancer are extremely promising, given the emerging evidence that supports therapeutic targeting of growth factors, growth-factor receptors and down-stream pathways [36]. Both bevacuzimab, a vascular endothelial growth factor antibody, and cetuximab, an endothelial growth factor-receptor antibody, have demonstrated benefit in advanced colorectal cancer [38, 39]. Oral fluoropyrimidine capecitabine has an improved safety profile and is at least as effective [40].

Palliative treatment for patients with metastatic colorectal cancer aims to improve survival and quality of life and major progress has been made by treatment regimens with new cytotoxic drugs such as irinotecan or oxaliplatin. Now the median overall survival for metastatic colon cancer has been doubled from 10 to 20 months [41].

The present state of generic therapy based on disease stage and histologic site of origin must give way to molecularly guided, personalized therapies, shifting the risk-benefit ratio of chemotherapeutic intervention towards a clear individualized patient benefit [36].

Follow-up

The main objective of follow-up after curative resection is improvement of survival but psychological support can be given as well. Meta-analyses showed a significant improvement in survival after intense compared with routine follow-up [42, 43]. However, how this should be done is not well defined. Colonoscopy is recommended every 3 years to detect metachronous colorectal cancer. Determination of carcinoembryonic-antigen testing every 3–6 months seems the most useful test in patients who had elevated CEA levels before tumour excision [44]. The value of other tests like abdominal ultrasound, CT scan and PET still needs to be determined [45].

When recurrent disease is detected at the previous site of resection, re-resection may be attempted. A selection of patients benefit from a combination of surgery, radiotherapy and/or chemotherapy. Repeat resections are performed for recurrent metastases after partial liver resection. Several series have shown survival curves after repeat resections of colorectal liver metastases, comparable to survival after first resection.

Conclusions

What do clinicians want to know from the radiologist? For population-based screening, further development of virtual colonoscopy is of great value. Preferably this is performed without bowel preparation and air-insufflation and with minimal exposure to radiation. Better imaging of flat and small adenomas and cancers should be made possible. Furthermore, molecular imaging has great potential for both depicting possible polyps and differentiating their nature (e.g. hyperplastic or adenomatous) and should be developed further.

For staging of colorectal cancer, which is the basis for therapeutic decisions, optimal imaging is necessary. The T-stage (with possible local invasion in adjacent structures or organs and/or tumour-related complications such as fistula formation or perforation), and possible lymph node metastases, hepatic metastases and pulmonary metastases should be assessed. Furthermore, the exact location in the colon should be determined and any second primary cancer or polyp be diagnosed. In the case of liver metastases, the number and location relative to the liver segments, possible involvement of major vessels (vena cava, portal vein, hepatic artery) and estimated volume of future remnant liver (when resection is planned) should be assessed.

In the follow-up after curative resection of colorectal cancer, early detection of recurrent disease, both at the site of the primary tumour and in the liver or lungs, might help to further improve survival and the value of the different imaging-modalities should be determined.