Figure 68-1. Trends in Colorectal Cancer Incidence and Death Rates by Sex, USA, 1930–2010. (From The American Cancer Society, Colorectal Cancer Facts and Figures 2014–2016.)
The list of behavioral or modifiable factors includes obesity, physical inactivity, smoking, diet, and alcohol intake.6 Obesity is associated with increased risk of developing CRC in both men and women.7,8 Abdominal obesity, measured as waist size, seems to be more important than excess weight, particularly in men. Being overweight increases the risk of CRC, independent of the level of physical activity. The most physically active people have a 25% lower risk of developing CRC, compared to the least active people. Both recreational and occupational physical activity reduces CRC risk, and sedentary people who become physically active later in life do reduce their CRC risk.9 The geographical differences in the incidence of CRC, and the change in incidence in migrant populations, suggest that diet is probably an important risk factor.10 The risk of CRC has been associated with a high consumption of red or processed meats.11,12 On the other hand, the risk of CRC is inversely correlated with the intake of fiber and whole grains, fruits and vegetable, dairy products and calcium, vitamin D, and folates.6 Alcohol consumption is causally associated with the development of CRC, and the effect seems to be dose-dependent. The risk seems to be higher for men than for women. There is no difference in the effect with regard to the type of alcohol.13 Smoking increases the incidence and mortality from CRC, and in particular, rectal cancer.14,15
Table 68-1 Factors Associated with Risk of Developing Colorectal Cancer
Three-quarters of CRCs occur in people who do not have any particular predisposing factors, and who are therefore considered to be at average risk. The remaining 25% of patients have hereditary risk factors that predispose them to the development of CRC: either a well-defined hereditary CRC syndrome (5%) or a close relative who has been diagnosed with the disease (20%). People with a first-degree relative with CRC have between a 1.9 and 4.4 relative risk of developing CRC, compared to the average population. The risk is higher when the affected relative was diagnosed at an early age, or when several relatives have been affected.16,17 CRC survivors have four times the lifetime risk of developing a new (metachronous) CRC, compared with people at average risk. The estimated mean annual incidence of metachronous CRC is 0.3%, with a cumulative incidence of 3.1% at 10 years.18 A personal history of adenomatous colorectal polyps also increases the risk of CRC, in particular for patients with larger or multiple polyps and an early age at diagnosis.19
Patients with chronic inflammatory bowel disease, both ulcerative colitis and Crohn disease, are at increased risk of developing CRC.20,21 The risk increases with the duration of the disease. It is estimated that, after 10 years of disease, the risk of developing CRC increases by around 1% each year. Thus, it is estimated that patients with 30 years of inflammatory bowel disease have an 18% risk of developing CRC. The risk is higher for patients diagnosed with inflammatory bowel disease at an early age, and with disease extending proximal to the splenic flexure. However, the incidence of CRC in patients with ulcerative colitis seems to be decreasing. A recent meta-analysis of population-based cohort studies reported a 1.7 increased risk of CRC among inflammatory bowel disease patients, compared to the general population, after adjusting for age, gender, and duration of disease.22 These changes are probably due to more effective treatments and the efficacy of surveillance programs.23
Figure 68-2. Estimated New Cancer Cases and Deaths Worldwide for Leading Cancer Sites by Sex and Level of Economic Development, 2012 (excluding nonmelanoma skin cancers). (From Torre LA, Bray F, Siegel RL, et al. Global Cancer Statistics, 2012. CA Cancer J Clin 2015;65:87–108.)
Figure 68-3. Mutation frequency across 224 human colorectal cancer samples from The Cancer Genome Atlas. Note the clear separation of hypermutated (median of 728 nonsilent mutations per tumor) and nonhypermutated (median of 58 nonsilent mutations). Red, MSI high, CIMP high, or MLH1 silenced; light blue, MSI low, or CIMP low; black, rectum; white, colon; gray, no data. (From Comprehensive Molecular Characterization of Human Colon and Rectal Cancer. The Cancer Genome Atlas Network. Nature 2012;487(7407):330–337.)
Patients with type II diabetes have a higher risk of developing CRC compared to the nondiabetic individual, even when adjusting for factors such as obesity and sedentary lifestyle.24 In addition, CRC survival rates are lower for diabetic patients compared to nondiabetic patients.25 The relationship between diabetes and CRC seems to be stronger for men than for women.
The most comprehensive molecular analysis of CRC to date has been conducted by The Cancer Genome Atlas Consortium.27 They have used state of the art technology to elucidate the mutational spectrum, the chromosomal and sub-chromosomal changes, the epigenetic regulation and transcriptional alterations in a large number of CRCs. A key finding that has helped frame our molecular understanding of CRC is the wide variation in the number of somatic mutations present in most tumors. According to the number of mutations, CRC can be effectively classified as hypermutated, harboring a median of 728 nonsilent mutations per tumor, and nonhypermutated, with a median of 58 nonsilent mutations (Fig. 68-3). Nonhypermutated tumors comprise 84% of CRCs, and are characterized by chromosomal instability (CIN). CIN is the result of missegregation of chromatids in tumor cells, resulting in gains and losses of entire chromosomes or chromosomal segments; manifesting as aneuploidy and copy number alterations (CNA).28 Common patterns of aneuploidy and CNAs in nonhypermutated CRC include amplifications in chromosome arms 1q, 7p and q, 8q, 13q, 17q, and 20p/q, and deletions in 8p, 14q, 15q, 17p, and 18p/q. Some of these CNAs could account for the loss of important tumor suppressors such as TP53 (on 17p), DCC, and SMAD4 (on 18q), and the amplification of oncogenes such as ERBB2 (on 17q). In short, CIN in nonhypermutated tumors can activate critical oncogenes, and inactivate tumor suppressors that contribute to the acquisition of tumorigenic traits.
The hypermutated tumors, representing 16% of all CRCs, are deficient in the mismatch-repair (MMR) mechanisms controlling the correction of errors that occur during DNA replication. The end result of MMR deficiency is a form of genetic instability characterized by the accumulation of mutations throughout the genome, primarily in repetitive sequences known as microsatellites; this is known as microsatellite instability (MSI).29 The mutational burden ultimately provides these MSI tumors with their malignant potential. Unlike CIN tumors, MSI tumors generally remain diploid or near-diploid. They are less likely to have KRAS, TP53, SMAD4 mutations, and more likely to have BRAF and TGFBR2 mutations (Fig. 68-4). MSI can arise from either the germline inactivating mutations or the epigenetic silencing of one of several MMR genes (MLH1, MLH3, MSH2, MSH3, and MSH6). Research into the mechanisms of epigenetic silencing of MSI tumors has helped elucidate the CpG Island Methylator Phenotype (CIMP), whereby the methylation of CpG sequences in the promoter region of MMR genes (most commonly MLH1) leads to their transcriptional silencing.
The classification of CRC according to the type of genomic instability is clinically relevant. MSI tumors tend to be preferentially right sided, have a solid or cribriform histologic pattern, contain large number of tumor-infiltrating lymphocytes, are less responsive to fluoropyrimidine, and in general carry a better prognosis, compared to CIN tumors.
The genomic instability that characterizes CRC is associated with alterations in a number of key pathways; among others, the WNT, MAPK, PI3K, TGF-β and p53 pathways (Fig. 68-5). The WNT signaling pathway contributes to the tightly regulated homeostasis of intestinal epithelial crypts, and its alteration is considered an initiating event in colorectal carcinogenesis. The WNT signaling pathway is altered in greater than 90% of both hypermutated and nonhypermutated CRCs. The most prevalent alteration leading to dysregulation of the WNT pathway occurs through biallelic inactivation of the tumor suppressor adenomatous polyposis coli (APC) gene.30 Patients with familial adenomatous polyposis (FAP) have a germline mutation in one allele of the APC gene, and their risk of developing CRC is virtually 100%. The WNT pathway is also commonly altered as a result of stabilizing mutations of the CTNNB1 gene coding for β-catenin, a protein normally targeted for degradation by APC. The accumulation of β-catenin as a result of APC loss or stabilizing CTNNB1 mutations leads to its translocation to the nucleus, where it interacts with transcription factors of the TCF/LCF family, turning them into transcriptional activators. The end result is an increase in the transcription of genes that are normally important for stem cell renewal and differentiation but, when inappropriately expressed at high levels, can cause cancer.
Figure 68-4. Most significantly mutated genes in hypermutated and nonhypermutated tumors. Blue bars represent genes identified using the MutSig algorithm, and black bars represent genes identified by manual examination of sequence data. (From Comprehensive Molecular Characterization of Human Colon and Rectal Cancer. The Cancer Genome Atlas Network. Nature 2012;487(7407):330–337.)
Mutational inactivation of the TGF-β pathway, important in regulating cell growth arrest and apoptosis, is a common event in CRC.31 In the majority of hypermutated tumors, the TGF-β gene is inactivated by a frameshift mutation in a polyadenine repeat within the coding sequence. It can also be inactivated by point mutation in the kinase domain. Downstream effectors of this pathway such as the transcription factor SMAD4, and associated proteins SMAD2 and SMAD3, are also inactivated by mutation or homozygous deletion of chromosomal segment 18q. Another member of this pathway, deleted in CRC (DCC), is also commonly inactivated by deletion of chromosome 18q.
Figure 68-5. Frequency of genetic changes leading to deregulation of recognized signaling pathways in CRC. Alterations are defined as somatic mutations, homozygous deletions, high-level focal amplifications, and in some cases, significant up- or downregulation of gene expression. Red/blue denote activation/inactivation, respectively. Bottom panel shows alterations across samples in five main pathways (WNT, TGF-B, RTK/KRAS, PI3K, TP53, if at least one gene in the pathway is altered). nHM, nonhypermutated; HM, hypermutated. (From Comprehensive Molecular Characterization of Human Colon and Rectal Cancer. The Cancer Genome Atlas Network. Nature 2012;487(7407):330–337.)
A common theme of the WNT and TGF-β signaling pathways is the increased activity of c-Myc, a regulator gene that codes for a multifunctional transcription factor which plays a role in cell cycle progression, apoptosis, and cellular transformation. c-Myc seems to play a central role in colorectal carcinogenesis.32
The TP53 gene is a member of a pathway of regulators of cell cycle arrest and cell death in response to a variety of genotoxic stresses. Mutations in the TP53 tumor suppressor gene occur in more than half of nonhypermutated CRCs. In most tumors, both alleles of the gene are inactivated by a combination of a mutation in one allele, and loss of the second allele by deletion of the chromosomal segment 17p. Loss of the TP53 tumor suppressor gene is generally considered an early event that plays a role in the transition from adenoma to invasive carcinoma.33 In hypermutated tumors, the TP53 pathway may be attenuated by mutations in other genes, such as the proapoptotic BAX.
The MAPK pathway consists of a cascade of tightly coordinate kinases that work together to regulate cell division and proliferation. KRAS, a protooncogene that becomes constitutively activated by mutation, resulting in uncontrolled cell proliferation, is mutated in 37% of CRC. Interestingly, a number of studies from model organisms reveal that KRAS mutation by itself is not sufficient to initiate tumorigenesis; in order to have oncogenic potential it requires a previous APC mutation. The MAPK pathway is the molecular pathway that has successfully been therapeutically targeted in CRC. Inhibition of this pathway using antibodies against the epidermal growth factor receptor (EGFR) reduces progression in wild-type KRAS/NRAS metastatic CRC.34 Activating mutations of BRAF, also an effector member of the MAPK pathway downstream of KRAS, are seen in 13% of CRC. The BRAF mutation, found almost exclusively in hypermutated tumors, has been associated with poor prognosis.
The PI3K/AKT/mTOR signaling pathway, with important roles in cell proliferation and apoptosis, is abnormally activated in nearly half of all CRCs. Moreover, in the TCGA cohort, PI3K and RAS pathways are simultaneously affected by mutations in one-third of all CRCs, suggesting that simultaneous inhibition of both pathways may be required to achieve a therapeutic effect.35
The characterization of genome-level alterations in CRC has led to the identification of new sub-classifications that share marked similarities across genomic structure, gene and protein expression, and even the activity of the tumor microenvironment.36–39 While these sub-classifications continue to be resolved and studied, they have shown potential to prognosticate tumor behavior and oncologic outcomes.40
The advances in genomic characterizations have revealed CRC to be a heterogeneous and complex disease. While the patterns of genomic instability and alterations of signaling pathways are aligned with specific phenotypic tumor characteristics, many tumors do not fit cleanly into any single category. Rather, they may host elements of genomic instability and dysregulated signaling pathways that may not be well characterized yet.
HISTOPATHOLOGY AND PROGRESSION
CRC is an adenocarcinoma arising from the epithelial lining of the large bowel. Some CRCs may develop de novo, but most result from malignant transformation of adenomatous polyps. In the past, only tubular and villous adenomas were considered to develop into invasive adenocarcinomas. However, recent evidence suggests that serrated polyps can also develop into CRCs. Polyps arise from normal mucosa, and gradually increase in size. Some polyp characteristics – size larger than 1 cm, tubulovillous or villous histology, multiple occurrences – are associated with a high risk of malignant transformation. Most screening programs are designed to recognize the presence of polyps, or early malignant change in polyps or the surrounding mucosa.
HISTOPATHOLOGIC TYPES OF COLORECTAL CANCER
Table 68-2 Colorectal Cancer: Histopathologic Types
The majority of CRCs are typically adenocarcinomas that form glandular structures resembling the normal colonic epithelium. However, there are several histologic types of adenocarcinoma (Table 68-2). Colorectal adenocarcinomas are assigned to one of four histologic grades according to their histologic resemblance to the normal colonic epithelium (Table 68-3). The histologic grade is associated with oncologic outcome independent of other risk factors, including tumor staging.
CRC is locally invasive, potentially spreading through the full thickness of the bowel wall into adjacent tissues. From the primary site, CRC often extends to the regional lymph nodes and to other organs. CRC can metastasize to almost any organ, but the most common sites are the liver and lungs. Approximately 20% to 34% of patients have metastases at the time of diagnosis, and another 30% of those initially treated with curative intent will subsequently develop distant metastases. Other sites of distant metastasis are the brain and bones, but these are unusual in the absence of liver metastases. Patients may also develop peritoneal spread, with the formation of malignant ascites. The risk of nodal metastasis increases with the depth of tumor invasion into the bowel wall, and the risk of distant metastasis is higher for patients with nodal metastasis.
STAGING AND PROGNOSIS
The anatomical extent or stage of the tumor at the time of diagnosis is a key factor in deciding upon treatment and defining prognosis in patients with CRC. The tumor–node–metastasis (TNM) classification developed by the UICC/AJCC is firmly established as the preferred staging system (Table 68-4). It classifies CRC according to the extent of the primary tumor (T), the involvement of the regional lymph nodes (N), and the presence or absence of distant metastasis (M). The TNM classification is important for both clinical and pathologic staging. Most colon cancers, and many rectal cancers, are staged after surgery and pathologic examination of the surgical specimen (pTNM). Many rectal cancer patients, and a small but growing number of colon cancer patients, do not receive surgery as the initial therapy, which makes clinical TNM (cTNM) staging based on medical history, physical examination, endoscopy, and imaging, increasingly important. As tumors experience a variable degree of regression with preoperative chemotherapy and/or radiation, it is important that patients receiving neoadjuvant therapy are assigned an accurate cTNM stage before starting treatment. For patients who have received neoadjuvant therapy, a modified pathologic staging is generated after surgical resection, annotated by the prefix y (ypTNM). The 7th edition of the AJCC Cancer Staging Manual incorporates a four-tier tumor regression grading (TRG) system for patients receiving neoadjuvant therapy. In this grading system TRG 0 represents Complete Response, no viable tumor cells or complete pathologic response (pCR); TRG 1, Moderate Response; TRG 2, Minimal Response; and TRG 3, Poor Response. Tumor regression seems to correlate closely with prognosis and long-term oncologic outcome.41
Table 68-3 Histologic Grading
The AJCC staging system is updated periodically in response to newly acquired clinical data and understanding of the biology of the tumor. The 7th edition of the AJCC Cancer Staging Manual, for all cancers diagnosed after January 1, 2010, was updated based on observed survival outcomes obtained from the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program.42 Relative survival in CRC patients decreases according to the depth of tumor penetration into the bowel wall and number of lymph nodes involved, but the presence of distant metastasis is the single most important predictor of survival in these patients (Table 68-4). Other tumor and treatment-related prognostic factors, in addition to the TNM stage, are listed in Table 68-5.
Quantitative parameters of lymph node evaluation – number of positive lymph nodes, total number of lymph nodes examined, and lymph node ratio – have prognostic implications in CRC. The number of nodes involved by tumor is included in the N category of the TNM system. The current staging system recommends examining a minimum of 12 lymph nodes for adequate staging.42 The total number of lymph nodes retrieved is lower in rectal cancer patients treated with preoperative chemotherapy and radiation, but the minimal number of nodes required for adequate staging in these patients is unknown. The total number of nodes examined has an important impact on outcome in patients with colon and rectal cancer; in patients with all T and N combinations, the probability of survival increases linearly with the number of lymph nodes examined. The lymph node ratio, defined as the number of positive lymph nodes divided by the total number of retrieved nodes, has also been considered an independent predictor of survival.43,44 However, the results have not been extensively validated and the lymph node ratio is not currently included in staging. Irregular tumor deposits in the mesocolon or mesorectum without surrounding lymph node tissue, are considered peritumoral tumor deposits; they are not counted as lymph nodes but are classified as N1c, and contribute to stage III.42 These are considered to represent large vessel or perineural invasion. The significance of micrometastasis detected by immunohistochemistry (usually with anticytokeratin antibodies) in H&E negative nodes of patients with stage II disease is controversial. Although some studies have reported an association between micrometastasis, increased recurrence and reduced survival, the detection of micrometastasis has not been incorporated into clinical practice, and at the present time should be considered investigational.45,46
The location of the tumor also provides prognostic information. In general, tumors located in the right colon have better prognosis compared with tumors located in the left or transverse colon. This may be attributed, in part, to the higher proportion of MSI tumors in the right side of the colon. In patients with rectal cancer, tumors located in the distal rectum have worse prognosis compared to tumors located in the upper rectum.
The quality of the surgery and the completeness of tumor resection also provide important prognostic information. The distance of tumor to the circumferential resection margin (CRM) – defined as the surgically dissected nonperitonealized surface of the specimen – provides important prognostic information in both colon and rectal cancers. The presence of tumor at the CRM, because of advanced stage or inadequate resection, is associated with a high rate of recurrence and decreased probability of survival. In rectal cancer patients, the CRM is defined as positive if the tumor, either by direct tumor extension or positive lymph node, is equal to or less than 1 mm from the resection margin.
The AJCC has established codes for completeness of the resection, which should be reported for each procedure.42 A resection is defined as R0 when the entire tumor is resected and the margins are histologically negative; R1 when the margins are grossly uninvolved but histologically positive; and R2 when the tumor is not completely resected, and there is gross residual tumor at the primary site, regional lymph nodes, or metastatic sites.
Other factors associated with worse outcomes include elevated CEA levels before surgery, high histologic grade, signet ring cell and small cell histopathologic type, lymphatic and blood vessel invasion, and perineural invasion.47–52 Vascular invasion is considered a sign of aggressive behavior in CRC, and has been associated with a higher risk of disease progression and poor prognosis. Capillary invasion is often associated with lymphatic invasion, and these are commonly designated as lymphovascular invasion. Large vessel invasion, detectable on magnetic resonance imaging (MRI), is now reported separately.
While the TNM staging system provides important prognostic information, it is inadequate for individual prognostication. Outcomes for individual patients within each tumor stage are variable. Recently, nomograms using a number of clinical and pathologic characteristics have been designed to improve prediction and survival beyond TNM staging in patients with colon or rectal cancer, particularly in the setting of stage II tumors (Fig. 68-6).53 In addition to providing prognostic information, nomograms may potentially guide clinical decisions, such as the use of adjuvant chemotherapy or the regularity of surveillance. However, nomograms have not become part of the treatment strategy.
A number of molecular markers such as MSI, 18q loss of heterozygosity, TP53, p21, among others, have been associated with prognosis in CRC patients. But none of these are routinely incorporated into clinical practice. Somatic RAS mutations have been associated with lack of response to anti–EGFR targeted therapies, and clinical guidelines worldwide now recommend testing for both KRAS and NRAS mutations if anti-EGFR therapy is contemplated in patients with stage IV disease.54 In addition, when a KRAS or NRAS mutation is not found, some recommend V600E BRAF testing because the data suggest that those tumors are also unresponsive to anti-EGFR therapy. However, these mutations have not been incorporated into the staging system.
Based on recent advances in the molecular characterization of CRC, a number of multigene molecular signatures have been designed as prognosticators of recurrence and outcomes. The Oncotype DX Colon Cancer Assay, developed by Genomic Health, Inc. (Redwood City, CA, USA) is an assay based on the expression of 12 genes that play roles in cell-cycle and stromal responses.55,56 The assay was designed and validated using formalin-fixed, paraffin-embedded tissue samples from patients enrolled in the Cancer and Leukemia Group B (CALGB) 9581 cohort.57,58 The predictive model provides a score predicting 3-year recurrence risk in stage II colon cancers, and appears most effective in MMR-proficient T3 tumors. Another commercially available tool is ColoPrint, developed by Agendia (Amsterdam, The Netherlands). With a similar premise, this tool is an 18-gene signature found to be predictive of distant metastasis in stage II colon cancers.59,60 ColDx, developed by Almac (Craigavon, UK) is a microarray-based tool that uses 634 probes to identify patients with stage II colon cancer at high risk of recurrence.61 While these molecular signatures seem to determine risk of recurrence independent of other well-established risk factors, they have not been fully incorporated into clinical practice.
Table 68-4 Clinical and Pathologic Staging4
Table 68-5 Colorectal Cancer Prognostic Factors
SCREENING AND PREVENTION
The slow development of CRC from a benign polyp to an invasive cancer provides a window of opportunity for the detection and removal of premalignant adenomatous polyps and early-stage cancers. Removal of adenomatous polyps reduces the incidence of cancer, and the diagnosis of CRC at earlier stages reduces mortality.62,63 A number of prospective studies have proven that screening for colorectal polyps and cancer using a variety of methods reduces CRC mortality, and the reduction in mortality persists long-term after screening.64–66 There is good evidence that screening has contributed significantly to the drop in CRC mortality rates from a peak a few decades ago. Screening for CRC is cost-effective, in terms of the quality-adjusted life-years gained, compared to nonscreening. In the United States, screening for CRC is recommended for men and women over age 50, but compliance remains suboptimal because more than one-third of Americans report not having participated in a screening program.
Figure 68-6. Nomogram to predict 5-year and 10-year recurrence-free survival in colon cancer. Used by drawing a straight line up to the Points axis to determine how many points toward recurrence the patient should receive. Sum of the points received from each prognostic variable is then located along the total points axis. Then drawing a line down from the total Points axis to the 5-year or 10-year freedom from recurrence axes provides the patient’s specific risk. RS, rectosigmoid colon; L, left colon; R, right colon; Sig, sigmoid colon; TC, transverse colon. (From Weiser et al. 2008, J Clin Oncol 26:380–385.)
Numerous screening methods for CRC have been used over the years (Tables 68-6 and 68-7). These fall into one of two categories: stool tests, which detect the presence of blood or altered DNA in the stool; and structural tests, which identify polyps and cancers. Efficacy in detecting CRC, cost-effectiveness as a screening tool, supporting evidence, and patient acceptability vary for each of these tests.67–69
CRCs and polyps bleed more than the normal mucosa, and detecting occult blood in the stool is the basis of the most widely used screening tests. Blood is detected by searching the stool for hemoglobin using chemical or immunologic methods; patients found to have blood in the stool should then undergo colonoscopy. The original fecal occult blood tests (FOBTs) relied on guaiac-based detection of the pseudoperoxidase activity of hemoglobin. However, as pseudoperoxidase activity is not specific to human hemoglobin, foods such as red meat can produce false positives. Medications such as aspirin and nonsteroidal anti-inflammatory drugs can also cause a false-positive reaction. Other foods, in particular those rich in vitamin C, can cause false-negative results and should also be avoided before a test. Thus, for improved accuracy, a special diet and avoidance of these drugs should be followed for 2 to 3 days before FOBT. As most tumors bleed slowly and intermittently, the sensitivity of this test remains low. Rehydration of the test cards increases sensitivity, at the cost of reducing specificity. The sensitivity of the test increases with the number of samples tested; testing two samples per stool on three consecutive bowel movements is recommended. Several prospective, randomized trials have demonstrated that screening by FOBT, followed by total colonic evaluation with colonoscopy in individuals with a positive test, reduces mortality from CRC.70–72
Fecal immunochemical tests (FITs) rely on antibodies that are specific to human hemoglobin, and the analysis of samples by automated quantification methods. FITs are as sensitive as the guaiac-based tests, but more specific in detecting human hemoglobin in stool. They therefore avoid the false-negative results in the presence of vitamin C, and the false positives obtained in guaiac-based testing from red meats. The test does not require dietary modification beforehand, and the handling of the specimens is less demanding. As with any fecal test, a positive result with FIT requires a complete colonoscopic examination. Several studies have demonstrated that FIT has better screening performance, compared to FOBT.73,74 As an additional consideration, FIT may be more easily implemented as a screening regimen, compared with sole usage of colonoscopy.75 Based on this evidence, a number of countries have introduced screening programs utilizing these tools.
Detection of altered DNA from exfoliated tumor cells has been investigated as a screening test for CRC for years. Similar to the detection of hemoglobin, detection of altered DNA triggers patient referral for colonoscopy. Large, prospective studies of this test show fair sensitivity in detecting CRC and low sensitivity in detecting large adenomas, compared with colonoscopy.76,77 A more recent study using new stabilizing buffers, more discriminating markers, and more sensitive analytical methods, has shown that stool DNA testing is more sensitive than FIT in detecting CRC and advanced precancerous lesions. However, the specificity of stool DNA testing was found to be inferior to FIT, with roughly 10% of the screened individuals having a false-positive result.78 There are concerns that a positive stool DNA test and negative colonoscopy may lead to additional and unnecessary work-up for malignancy. Therefore, screening guidelines of the U.S. Preventive Services Task Force do not currently recommend the fecal DNA test as a screening option.
Two-thirds of CRCs and polyps are located in the sigmoid colon and rectum, and can be reached with a 60-cm flexible sigmoidoscope. The presence of adenomatous polyps in the rectosigmoid colon increases the probability of finding additional polyps or cancers in more proximal segments of the large bowel. If an adenomatous polyp is found during flexible sigmoidoscopy, the patient should undergo a complete colonoscopy. Flexible sigmoidoscopy is safe, fast, requires minimal preparation, and can be performed in an office-based setting, as conscious sedation is not needed. The risk of perforation with flexible sigmoidoscopy is approximately 1 in 20,000, but the lack of sedation can occasionally be associated with discomfort, deterring some patients from undergoing future examinations. The effectiveness of flexible sigmoidoscopy as a screening modality requires examination to at least 40 cm from the anal verge, and the ability of the endoscopist to biopsy-suspected adenomas. The main limitation of flexible sigmoidoscopy is that it does not examine the entire colon. However, as distal tubular adenomas are often indicative of proximal advanced neoplasia, the efficacy of flexible sigmoidoscopy is greatest when patients with distal adenomas are subsequently referred for colonoscopy. Due to differences in the distribution of colorectal neoplasia in patients of different age, gender, and racial groups, the efficacy of flexible sigmoidoscopy may vary. Several case-control studies have demonstrated that screening by sigmoidoscopy reduces mortality from CRC by two-thirds in the setting of tumors located within reach of the sigmoidoscope.79,80 More recently, several prospective studies have demonstrated that screening with flexible sigmoidoscopy reduced CRC incidence and mortality by approximately 25%.81–83 The reduction in incidence occurs in both the proximal and distal colon, while the reduction in mortality applies mainly to tumors in the distal colon. The optimal interval between tests is still controversial. In some studies, flexible sigmoidoscopy was performed every 3 to 5 years. At least two prospective trials demonstrated a reduction in CRC incidence and mortality with only one flexible sigmoidoscopy screening between 55 and 64 years of age.83,84
Although the evidence for combining FOBT and flexible sigmoidoscopy is weak, some studies have shown that the combination of these two screening methods is more effective in detecting colorectal neoplasia than each method used individually.85 The combined approach has a theoretical advantage of detecting lesions located throughout the colon, but its impact on mortality from CRC is unknown. In the United States, annual FOBT combined with flexible sigmoidoscopy every 5 years is a common screening method for the average-risk population.
Colonoscopy is considered the most accurate test for the early diagnosis and prevention of CRC. It allows direct visualization of the mucosa of the entire colon and rectum, simultaneously allowing the biopsy or removal of suspicious lesions. Colonoscopy is also used to evaluate patients who have tested positive on other screening tests. However, colonoscopy is inconvenient, requires dietary modification and bowel preparation beforehand, is usually performed under conscious sedation, and carries a risk of complications of 1–2 per 1,000. Colonoscopy is also more expensive compared to other screening methods. Overall, patient acceptability of colonoscopy seems to be higher compared to other invasive screening methods, and it is now the most commonly used screening method in the United States. There is indirect evidence from microsimulation and case-control studies that colonoscopy reduces mortality.86 However, randomized controlled trials proving that screening with colonoscopy reduces CRC mortality are lacking. Comparative studies show that colonoscopy is more effective in detecting advanced colonic neoplasia, in both men and women, than a single FOBT combined with sigmoidoscopy. Colonoscopy is more likely to detect preneoplastic polyps, but just as likely to detect invasive CRC as FITs. In addition, colonoscopy provides the protective benefit of screening the proximal colon.87 For average-risk individuals colonoscopy screening should start at 50 years of age and be repeated every 10 years.
Table 68-7 Screening and Surveillance for Colorectal Cancer According to Risk7
Double-contrast barium enema can detect most of the clinically important lesions in the colon but, as with colonoscopy, its effectiveness as a screening test for CRC is based only on indirect evidence. Similar to colonoscopy, it requires dietary modification and mechanical bowel preparation, and is associated with patient discomfort. In addition, a positive test mandates a colonoscopy. While double-contrast barium enema every 5 years should provide the same degree of protection as the other screening strategies, it is rarely used as a primary screening method today. Its clinical role as a screening tool at this time is primarily for visualization of the colon in patients who cannot undergo a complete colonoscopy.
Computed tomography colonography (CTC), also referred to as virtual colonoscopy, involves thin-section, multidetector, helical CT, and three-dimensional viewing for interpretation.88 CTC identifies 90% of cancers or adenomas measuring more than 10 mm in diameter in asymptomatic individuals 50 years of age or older.89 It is more sensitive than other screening methods, but can miss flat lesions and polyps smaller than 10 mm in diameter. CTC obviates some of the drawbacks of colonoscopy, such as the need for sedation and recovery time, but also has downstream consequences: it delivers a dose of radiation that may become substantial with repeated examinations, and detects incidental extra-colonic findings that may trigger expensive, and sometimes unnecessary diagnostic investigations. Furthermore, it requires standard bowel preparation and gaseous distension of the colon. New methods of tagging and subtracting residual stool may obviate the routine need for mechanical bowel preparation. Finally, patients with lesions found on CTC still require full visualization of the colon by colonoscopy. CTC is now considered an acceptable screening alternative for average-risk individuals, starting at 50 years of age; the interval between such examinations remains uncertain, although a 5-year interval has been suggested, based on computer simulation models.90
Patients with symptoms of CRC should undergo the appropriate diagnostic studies; they are not candidates for screening. Screening recommendations for the general population are based on individual risk assessment. Based on the past medical history and family history of CRC or polyps, individuals are assigned to one of three risk categories, with different screening recommendations (Table 68-7).91
People at average risk for the development of CRC (asymptomatic men and women without risk factors, over 50 years of age in the United States and 60 in the United Kingdom) could undergo yearly FOBT, combined with flexible sigmoidoscopy every 5 years. People with a positive FOBT or a polyp identified by flexible sigmoidoscopy should undergo entire colon and rectum examination by colonoscopy. Double-contrast barium enema every 5 years, or colonoscopy every 10 years, is an accepted screening alternative in the average-risk population. CTC every 5 years is also an option. Digital rectal examination (DRE) should be performed at the time of sigmoidoscopy or colonoscopy in all individuals.
Family History. Individuals with a first-degree relative (parent, sibling, or child) with CRC or adenomatous polyps diagnosed before 60 years of age should start screening with colonoscopy at 40 years of age or 10 years younger than the earliest diagnosis in their family, whichever comes first. The test should be repeated every 5 years. Individuals with one first-degree relative with CRC or adenomatous polyp diagnosed at 60 years of age or older, or with two or more second-degree relatives (grandparent, grandchild, aunt, uncle, niece, nephew, half-sibling) with CRC or adenomatous polyps at any age, should undergo screening as average-risk individuals – but starting at 40 years of age. Patients with one second-degree relative or one or more third-degree relatives (first cousin) affected are considered to be at average risk.
History of Polyps at Previous Colonoscopy. Patients undergoing endoscopic excision of a small (<1 cm) adenomatous polyp should have the entire colon examined at the time of the polypectomy. Colonoscopy should be repeated 5 years later. If the test is negative, they should then follow the screening recommendations for average risk.92
Patients with more than three adenomas, adenomas with villous features or high-grade dysplasia, or a large adenomatous polyp (>1 cm) who have the entire colon examined at the time of polypectomy, should undergo complete colonoscopy 3 years later – and, if normal, every 5 years thereafter.
A complete colonic examination should be carried out before surgery in any patient undergoing a planned curative resection for CRC. The colonoscopy should be repeated at 1 year to exclude metachronous lesions. If the examination at 1 year is normal, it should be repeated after 3 years, and every 5 years thereafter if the previous one was normal.
Patients with Colorectal Cancer. Patients with colon and rectal cancer should undergo high-quality perioperative clearing of polyps with colonoscopy 3 to 6 months after resection, if not performed before surgery. Surveillance after curative resection is described in detail in Surveillance after Curative Resection for Colon and Rectal Cancer section.
This category includes individuals from families diagnosed with hereditary forms of CRC.
Familial Adenomatous Polyposis. Once the diagnosis of FAP is established, the patient should undergo colectomy, or a yearly colonoscopy until colectomy. Upper gastrointestinal endoscopy should be performed every 1 to 2 years. Siblings and children of a patient with FAP should start surveillance by flexible sigmoidoscopy at puberty.
MYH-Associated Polyposis. Depending on the individual, age of presentation, and number and size of polyps, the patient may be advised to undergo a prophylactic colectomy or yearly colonoscopy beginning at 25 years of age. Upper gastrointestinal surveillance should be performed following the guidelines for patients with FAP.
Lynch Syndrome. Individuals from families fitting the Amsterdam criteria for HNPCC should have a colonoscopy at 21 years of age, then every 2 years until 40 years of age, and yearly thereafter. Genetic counseling and testing should be considered.
At-risk members of families with hereditary cancer syndromes should be informed about the benefits and limitations of genetic counseling and genetic testing.
Consequences of Screening
The full spectrum of clinical consequences of screening, other than the prevention of deaths from CRC, is difficult to predict because every screening strategy initiates a cascade of events, each one with uncertain probability. The rate of false-positive tests, the number of colonoscopies performed, the complications of the screening and diagnostic tests, and the number of patients who may require surveillance as a consequence of screening, are complex factors that arise at different times over several decades. However, cost-effectiveness analysis in the United States has demonstrated that screening for CRC in average-risk patients, according to the strategies outlined above, compares favorably with other healthcare interventions such as mammography or treatment of mild hypertension.93,94 It is important to understand that screening does not completely eliminate the risk of CRC. Up to 9% of CRCs are diagnosed within 6 to 36 months after a screening colonoscopy.95,96 These interval cancers tend to be preferentially located in the proximal colon. This may be related to failure to reach the proximal colon, residual stool obscuring the view in the proximal colon, or different characteristics of the proximal lesions that makes them more difficult to detect or remove.
Implementation of Guidelines
The general population has limited awareness of the risks of CRC or its symptoms; as a result the number of individuals participating in screening programs has been low. The dissemination of information to patients is an essential part of the screening program. Primary care physicians have a responsibility to inform their patients about their risk of CRC, the benefits of screening and different screening strategies, and to set up a system for implementing these guidelines.94,97,98 Through such efforts, the proportion of adults 50 to 75 years of age having colonoscopy has increased steadily in the last decade: from 19.1% in 2000 to 55% in 2010.3
While the proportion of CRCs diagnosed through screening is increasing, most are still diagnosed after patients become symptomatic. CRC can cause a myriad of symptoms, many of which are nonspecific and highly prevalent among healthy individuals. Therefore, the diagnosis of colorectal carcinoma often requires a high index of suspicion.
Symptoms and Signs
The most common symptoms in patients with CRC are abdominal pain, change in bowel habits, rectal bleeding, anemia, anorexia, and weight loss. The clinical manifestations of CRC differ depending on the location and stage of the tumor.
Abdominal pain is nonspecific; it may be localized to any quadrant of the abdomen, or may be diffuse. When the pain is persistent and colicky, it is more likely to represent obstructive symptoms resulting from a left-sided lesion. More localized tenderness with signs of localized peritonitis indicates local invasion of the adjacent peritoneum or perforation. It can be difficult to distinguish the pain associated with diverticular disease from that due to a carcinoma in the sigmoid or descending colon. Patients with diverticular disease may also have an underlying carcinoma. Abdominal pain may occasionally be related to extensive metastatic disease in the liver or retroperitoneal lymph nodes. Rectal cancer can also cause perineal discomfort and is often associated with tenesmus, the sensation of incomplete defecation. A locally advanced rectal cancer can also cause sacral or sciatic pain.
An unexplained change in bowel habits lasting more than 2 weeks requires investigation for CRC. Constipation associated with colicky abdominal pain is a common manifestation of partially obstructing tumors. Abdominal pain, constipation, and abdominal distension suggest a complete colonic obstruction, which is more common in tumors located in the sigmoid colon and rectum. Some patients with sigmoid or rectal cancer may also complain of diarrhea, which may sometimes be bloody.
Bleeding from tumors in the sigmoid colon and rectum is often wrongly attributed to hemorrhoids or other benign anal conditions. However, blood from hemorrhoids is usually bright red and is accompanied by anal discomfort; the bleeding is intermittent and splashes the toilet pan. Anal fissure is also a common cause of rectal bleeding; the blood is often bright, occurring after defecation, and is typically associated with sharp anal pain triggered by defecation. Painless defecation with rectal blood that is darker in color and mixed in with stool is more likely to be secondary to an underlying carcinoma. Rectal bleeding, particularly when associated with tenesmus, requires a thorough diagnostic investigation to exclude an underlying rectal tumor.
The development of nonspecific anemia of unknown origin is a common manifestation in patients with a carcinoma in the proximal colon. Patients with unexplained microcytic anemia should be examined initially with colonoscopy. Anorexia and weight loss frequently accompany CRC, and are often associated with advanced disease. The differential diagnosis in these patients is a gastric carcinoma, but when investigations are negative it is important to exclude a large-bowel malignancy. As the bowel lumen is larger and the stools overall looser in the proximal colon, obstructive symptoms are uncommon; therefore, patients with tumors proximal to the splenic flexure often present with advanced disease.
It should be remembered that the presence of colon cancer can also be identified incidentally during investigations for other pathologies such as gallstones, gynecologic, or urinary conditions. A complete family history, with emphasis on past history of CRC or polyps, is essential to diagnose some hereditary cancer syndromes. A family history of endometrial, gastric, urologic, or other HNPCC-associated cancers is also important because it may help in diagnosing Lynch II syndrome.
The evaluation of patients suspected of CRC requires a complete physical examination. Features that should arouse suspicion of malignancy include pallor, palpable abdominal mass, and a palpable mass on DRE. Hepatomegaly indicates advanced disease with extensive liver metastases, and consequently a very poor prognosis. A hard mass in the pouch of Douglas felt on DRE may indicate peritoneal carcinomatosis. Other signs of an underlying CRC include pneumaturia, ischiorectal or perineal abscesses, or even deep venous thrombosis.
The differential diagnosis of CRC includes diverticulitis, irritable bowel syndrome, inflammatory bowel disease, ischemic colitis, and benign anorectal conditions such as hemorrhoids or rectal mucosal prolapse.
Patients with symptoms suggesting CRC should have a colonoscopy to evaluate the entire colon (Fig. 68-7). The examination should be complete, as up to 4% of patients with CRC have synchronous cancers and many more have colorectal polyps.99,100 Up to one-third of these synchronous cancers are in locations requiring a different surgical resection, further emphasizing the importance of complete colonoscopy. As many patients undergo minimally invasive surgery today, the endoscopist should mark the vicinity of the tumor with India ink to help locate the lesion intraoperatively. CT colonography and DCBE are less effective in investigating patients with symptoms suggestive of CRC, but they are useful in patients with partially obstructive tumors, in whom colonoscopy cannot be completed. In patients with obstructive lesions, the proximal colon can be examined either by preoperative imaging studies such as PET-CT, or intraoperatively by direct palpation of the colon. In any case, patients with a complete colonoscopy before surgery should have a surveillance colonoscopy 6 months after surgery.
Assessment of the extent of disease at the time of diagnosis is important because clinical stage, particularly in rectal cancer, dictates treatment decisions. A CT scan of the chest, abdomen, and pelvis with intravenous and oral contrast is important in order to locate the tumor, detect mesenteric nodes, and exclude liver metastasis and gross peritoneal carcinomatosis (Fig. 68-7). In addition, preoperative imaging helps the surgeon plan the resection. In patients with iodine allergy, and in patients with undetermined lesions on CT, a PET-CT or and abdominal or pelvic MRI may provide additional information.
Laboratory evaluation includes complete blood cell count, coagulation parameters, and chemistry panel. Preoperative CEA should be measured because it provides prognostic information, and can be useful during patient surveillance.
Surgery is the primary treatment for CRC, but for many patients surgery is not the initial form of treatment and some do not require surgery at all. Treatment is different for colon than for rectal cancer, and depends also on the clinical stage of the tumor. Treatment decisions must also take into account the patient’s performance status, comorbidities, hereditary CRC predisposition, as well as desires and expectations.
All patients undergoing surgery for colon and rectal cancer require similar preoperative preparation aimed at optimizing the technical success of the procedure and avoiding perioperative complications. Surgical site infection (SSI) is the most common complication after colon and rectal cancer resection, occurring in 4% to 26% of cases.101,102 In addition to causing patient suffering, SSIs lengthen hospital stay and increase costs.103 The institutional rate of SSI after colorectal surgery is now an important quality metric for several benchmarking programs. A number of federal and state-wide initiatives aimed at improving surgical care recommend implementing measures to reduce the rate of SSI and other complications in patients undergoing elective colorectal surgery.104,105
Figure 68-7. Staging modalities for colorectal cancer. A: Obstructing colon cancer at colonoscopy. B: Polypoid colonic tumor at colonoscopy. C: Cecal circumferential tumor with regional lymphadenopathies. D: Liver metastases of colorectal origin.
The colon has the largest concentration of bacteria in the human body. A gram of feces contains 1011 polymicrobial bacteria, mostly gram negatives and anaerobes.106 Oral mechanical bowel preparation using polyethylene glycol to reduce the bacterial load and risk of intraoperative fecal spillage has been considered an axiom in colon and rectal surgery. However a number of prospective trials have failed to demonstrate benefit from mechanical bowel cleansing in preventing SSIs.107,108 These results were confirmed by a Cochrane systematic review of 5,805 patients; the authors concluded that there is no statistically significant evidence that patients benefit from mechanical bowel preparation or the use of rectal enemas.109 Another recent systematic review by the Agency for Health Care Research and Quality reached similar conclusions. Oral mechanical bowel preparation appeared to be protective, compared to no preparation, for peritonitis or intra-abdominal abscess, but the evidence was weak. The study could not draw any conclusion on potential harms, such as dehydration and electrolyte imbalances, related to use of oral mechanical bowel preparation.110 Despite the lack of solid data, many surgeons still recommend oral mechanical bowel preparation because manipulation and suturing is easier with a clean colon.
High-quality evidence indicates that antibiotics covering aerobic and anaerobic bacteria, delivered orally or intravenously (or both) prior to elective colorectal surgery, reduce the risk of postoperative surgical wound infection by as much as 66%.111 Oral antibiotics, neomycin- and erythromycin-based, are delivered the day before surgery, in combination with the oral mechanical bowel preparation. For patients without penicillin allergy, a second-generation cephalosporin (cefotetan or cefoxitin) is administered intravenously within 30 minutes of the surgical incision, with re-dosing during the procedure as required according to the half-life of the drug and the duration of surgery. For penicillin-allergic patients, metronidazole or clindamycin combined with either ciprofloxacin or gentamicin is acceptable, as are aztreonam and fluoroquinolones.112 Ertapenem, a long-acting carbapenem active against gram negatives and anaerobes, is an accepted alternative to second-generation cephalosporins for prophylaxis in CRC. Ertapenem has a long life and does not require re-dosing in prolonged procedures, but has been associated with increasing risk of Clostridium difficile infection.113 Other measures that prevent SSI include tight glucose control in diabetic patients, smoking cessation, clipping rather than shaving the skin of the abdominal wall, and maintaining normothermia and adequate oxygenation during anesthesia.114
Patients undergoing surgery for CRC are also at risk of deep venous thrombosis and pulmonary embolism, and should have thromboembolic prophylaxis with unfractionated heparin or low-molecular-weight heparin during the peri- and postoperative periods.115
As the incidence of CRC increases with age, many patients have cardiovascular or respiratory conditions requiring medical clearance before surgery. While technical advances have made CRC operations safer, optimal outcomes require special effort to ensure that the patient’s overall health is optimal at the time of surgery. Many CRC patients have other comorbid conditions such as diabetes, hypertension, and coronary artery disease requiring medical evaluation before undergoing surgery. Comorbidities can impact decision-making and affect short- and long-term outcomes. Patient clinical and performance status should be optimized to reduce the risk of perioperative complications. Fertility options should be discussed with all individuals of child-bearing potential.
Patients who may require a stoma should be seen before surgery by an enterostomal therapist. Adequate marking of the site improves outcomes for patients requiring a stoma. Preoperative teaching shortens the time patients require to become proficient managing the stoma, and reduces hospital stay.116
The Enhanced Recovery After Surgery (ERAS) protocols were introduced in open colorectal surgery in the 1990s, with the aim of speeding patient recovery, improving patient outcomes and satisfaction, shortening hospitalization, and reducing healthcare costs.117 ERAS protocols span the entire perioperative period, and attempt to minimize surgical stress and postoperative ileus through patient education, preoperative hydration and carbohydrate loading, goal-directed intraoperative fluid management, narcotic sparing for intraoperative and postoperative pain control, and early mobilization and oral feeding in the postoperative period. A number of prospective trials have indicated that the implementation of ERAS protocols reduces length of hospital stay, compared to conventional recovery in patients undergoing open or minimally invasive surgery for CRC.118,119 A recent systematic review and a meta-analysis confirmed that ERAS protocols resulted in a shorter length of stay and a reduction in overall complications, with no difference in mortality and surgical complications.120,121 Similar results have recently been reported from an international registry.122
Principles of Surgical Treatment
The goal for any curative-intent surgery is to remove the tumor-bearing segment of the bowel with adequate margins, along with en bloc excision of the mesentery containing the feeding vessels and regional lymph nodes. The location of the primary tumor determines lymphatic drainage and dictates the extent of the resection. Lymphatic capillaries are primarily located in the submucosal and subserosal layers of the bowel wall. The lymphatic flow in the colon is primarily circumferential, with longitudinal spread along the bowel wall thought to be less than 1 cm in each direction. Therefore, a 5-cm margin of normal bowel on either side of the primary tumor is considered sufficient to avoid anastomotic recurrence. The length of the terminal ileum resected in patients with rectal cancer does not influence the risk of anastomotic recurrence. In the rectum, where the longitudinal lymphatic flow is primarily upward, cancer cells rarely spread distally along the bowel wall farther than 1 cm from the macroscopic distal end of the tumor. Consequently, a 2-cm margin of normal bowel distal to the tumor, or even less in patients treated with neoadjuvant therapy, is considered appropriate to an oncologically safe resection.123
The lymphatic channels in the bowel wall drain to the regional nodes, which are classified in different groups according to their proximity to the bowel and its blood supply (Fig. 68-8). The “epicolic” nodes are located in the bowel wall under the peritoneum, usually close to the epiploic appendices. The “paracolic” lymph nodes are located along the marginal vessels. Next, the “intermediate” nodes are positioned in the middle of the mesentery. Finally, the “central” or “apical” lymph nodes are located close to the root of the mesentery, near the origin of the named vessels. While CRC generally spreads sequentially from the paracolic to the central or apical lymph nodes, nodal metastases skipping one of the groups are common.124
The extent of the mesenteric resection is determined by the need to remove all the lymph nodes draining the corresponding segment of bowel, including the central lymph nodes located at the origin of the named feeding blood vessels. As most tumors are located between two named vascular pedicles, both of these should be resected at their origin. When suspected of being involved by tumor, the central nodes should be marked on the specimen, as they have negative prognostic information. Other lymph nodes located away from the feeding vessels and suspected of tumor involvement during surgery, should be removed and analyzed, because in some patients the lymphatic drainage does not follow an orderly pattern.125,126 If residual metastatic lymph nodes remain after sampling, the resection should be considered incomplete.
Figure 68-8. Anatomy of the colon. A: Epicolic lymph nodes. B: Paracolic lymph nodes. C: Intermediate lymph nodes. D: Apical-central lymph nodes.