Polyposis syndromes: Familial adenomatous polyposis (FAP) (p 146) represents less than 1% of the total cases, and those with the FAP gene have a near 100% rate of cancer development. Pts with the HNPCC family hx represent about 1-2% of total cases and have a lifetime risk of CRC of about 80% if they inherit an HNPCC gene (p 144). Peutz-Jeghers (p 149), hyperplastic polyposis (p 151), and juvenile polyposis (p 150) are uncommon risk factors.

IBD: Those with ulcerative colitis or Crohn’s colitis have an increased rate of CRC after 8-10 yr, and this risk is 7-14% at 25 yr (p 114).

Family hx of CRC: Having a single first-degree relative with CRC appears to confer a lifetime risk of 12-24%. The high-risk end of the range represents people with a firstdegree relative with CRC under the age of 45 yr (Gastrointest Endosc Clin N Am 1993;3:715;
Ann IM 1993;118:785). Having 2 first-degree relatives with CRC puts the lifetime risk at 25-35%. The period of greatest excess RR begins under age 45 yr (when rates in the general population are low) (Nejm 1994;331:1669). Therefore, screening such individuals has maximum potential benefit if begun early (eg, age 40 yr). When only a second- or third-degree relative is affected, the risk is increased to only 30-50% above that of the general population, so the screening strategy is notaltered (Gastroenterol Clin North Am 1996;25:793).

Family hx of adenoma: Data from the National Cooperative Polyp Study suggests that the RR for CRC for pts whose first-degree relatives have an adenoma is 1.78, and this climbs to 2.59 for siblings of pts whose adenomas were diagnosed before age 60 yr (Nejm 1996;334:82). It has been recently argued that these data are really looking at the risk of adenomas in pts with CRC (Am J Gastro 2009;104:739).

Prior CRC: The risk of a second cancer is low in the first 10 yr after dx but may be as high as 6.3% at 18 yr (Clin Radiol 1984;35:425).

Prior colon adenoma: Pts with villous or tubulovillous adenomas >1 cm have a RR for cancer of 3.6 at a mean of 14 yr of follow-up (Nejm 1992;326:658).

Serrated colonic polyps: Most serrated polyps of the colon are hyperplastic polyps, which have conventionally been regarded as harmless. However, recent lines of evidence suggest that notall serrated polyps of the colon have a benign behavior. The “sessile serrated adenoma” is a serrated polyp that has architectural dysplasia rather than cytologic dysplasia (Am J Clin Pathol 2005;124:380). The polyps are commonly misdiagnosed as hyperplastic polyps. They typically occur in the right colon and can be large. They show a high frequency of the type of mutations found in colon cancers characterized by methylation of DNA (GE 2006;131:1400).

Other risks: Type 2 diabetes mellitus is associated with a RR = 1.43 for CRC in women (J Natl Cancer Inst 1999;91:542). Obesity is associated with a RR = 1.5-2.8 (Am J Gastro 2009;104:739). Cigarette smoking is an often overlooked risk (J Natl Cancer Inst 2000;92:1178). Other risks include prior ureterosigmoidostomy (RR = 10) (J Urol 1990;144:1110), irradiation for gynecologic cancer, acromegaly (Clin Endocrinol [Oxf] 1990;32:65), high intake of red meat (Cancer Res 1994;54:2390), and dietary fat (Nejm 1990;323:1664). Population-based cohort studies do notshow a link between breast cancer and CRC, though less methodologically sound studies have suggested an association (J Clin Gastroenterol 1994;19:57).

Protective factors: It has long been thought that dietary fiber is protective for CRC after Burkitt hypothesized that the low incidence of CRC in black Africans was related to high dietary fiber. The low risk in black Africans nowappears due to low animal product consumption (Am J Gastro 1999;94:1373). The data are notstriking for the protective effect of fiber. A large meta-analysis showed benefit (J Natl Cancer Inst 1990;82:650), but several cohort studies show no benefit (Gastroenterol Clin North Am 1996;25:717). Current thinking is that fruit and vegetable fiber are more important than grain fiber. The Nurses Health Study showed a clear benefit of >15 yr of folate-containing multivitamin use (RR = 0.25) and a more modest protective effect of dietary folate (Ann IM 1998;129:517).

The progression from adenoma to carcinoma: The molecular mechanisms that cause the formation of adenomas and their degeneration into cancers are quite well understood (Gut 1993;34:289). The common first step is mutation of the adenomatous polyposis coli (APC) gene. This gene (an inherited mutation of which is the defect in FAP [p 146]) appears to be critical in regulating cell proliferation and programmed cell death (apoptosis) (Curr Gastroenterol Rep 1999;1:449). Defects in the APC gene might be inherited or acquired. For example, a subtle mutation in the APC gene (the I1307K mutation) is responsible for inherited cancers in Ashkenazi Jews (Nat Genet 1997;17:79). Adenomas begin as aberrant crypt foci, and their subsequent growth from small adenoma to large adenoma is marked by a series of mutations. Many of these mutations occur in DNA mismatch repair genes (the described defects in HNPCC pts [p 144]). One described defect is the activation of an oncogene, K-ras, while the other defects are primarily loss of tumor suppressor genes. Further transformation to cancer occurs with mutation of the p53 gene, and metastatic disease arises by a series of additional mutations. This sequence does notoccur uniformly, and additional mutations are sure to be identified. The time required for this transformation from adenoma to cancer by multiple mutations cannot be precisely known. In the National Cooperative Polyp Study, only 5 of 1418 pts with adenomas developed cancers in 6 yr; all were early stage. Using available data on incidence and prevalence of polyps and data on the average pt age difference at dx of adenoma and cancer, an expert panel estimated that a polyp probably takes an average of 10 yr to transform into a malignancy (GE 1997;112:594). This allows ample chance to interrupt the process by endoscopic polypectomy.

Cancers without adenoma: A small proportion of CRCs appear to arise without a preexisting adenoma. In their early stages, these lesions are entirely flat or minimally elevated, often with a depressed center (Gastrointest Endosc 1995;41:135), and they appear to be capable of invading submucosa. It is difficult to estimate the proportion of cancers that come from such lesions.

Cancers from serrated polyps: About 15% of colon cancers show high levels of methylation of DNA. Current evidence suggests that serrated polyps, rather than adenomas, are the precursors for this type of colon cancer (GE 2006;131:1400).

Synchronous cancers (cancers diagnosed at the same time as the first cancer found) occur in 2-5% of pts, and their dx affects the extent of operation about 10% of the time (Surgery 1997;122:706).

Metachronous lesions (lesions diagnosed later in follow-up) occur in about 2% of pts with a mean lag of 9 yr (Surgery 1997;122:706). The rate of metachronous lesions is higher in pts with HNPCC and in pts who present with synchronous lesions (Dis Colon Rectum 1997;40:935).

Location of cancers: There has been an unexplained trend over time for CRC to be found proximal to the sigmoid colon (Jama 1977;238:1641).

Diet and chemoprevention: It is reasonable to recommend to pts a diet rich in fruit and vegetable fiber and low in fat and animal products; such a diet appears protective for CRC and other illnesses. Diets rich in cereal grains are more useful for treating constipation and preventing diverticular disease but could still be advocated as means of achieving a low-fat intake and for possible beneficial effects in CRC prevention. The data for aspirin are notcompelling enough to recommend to the population at large for CRC prevention given the associated risks. Assuming a 50% effectiveness for aspirin in risk reduction, its use would save fewer lives at greater cost than screening colonoscopy (GE 2002;122:78). Further study is needed to define the role of chemoprevention by use of calcium supplements, aspirin, NSAIDs, and other agents (reviewed in GE 2004;126:1423).

Detection and removal of adenomas (see also p 141): It is nowwell established that the most effective means of preventing CRC is to detect and remove adenomas.
The most convincing data are from the National Cooperative Polyp Study cohort. These 1418 pts had 1 or more adenomas removed and were followed for an average of 5.9 yr. When compared to 3 historic reference groups, the incidence of cancer was 76-90% lower than expected (Nejm 1993;329:1977). Case control studies of screening rigid sigmoidoscopy in a large health plan (Nejm 1992;326:653) and endoscopy with polypectomy in U.S. veterans (Ann IM 1995;123:904) demonstrated a 50% reduction in risk in the screened portions of bowel.

Surgery: All stage I-III cancers should be treated surgically for an attempt at cure. Stage IV lesions are treated surgically to palliate pain, bleeding, or obstruction. The resection involves the primary tumor and its lymphatic drainage with a 5-cm proximal margin and 2-cm distal margin of normal bowel. The extent of bowel resected is usually dictated by the vascular supply of the involved segment. Laparoscopic colectomy improves the immediate postoperative course compared to the open operation and is just as effective for long-term cancer cure as the open operation (Lancet 2004;363:1187; Nejm 2004;350:2050). Pts who present with complete obstruction require a 2-stage operation. They initially undergo diverting loop colostomy (in which a loop of colon is brought out to the skin and an ostomy with 2 barrels is created) until the bowel can be prepped and a cancer operation done.

Rectal cancers (Gastroenterol Clin North Am 1997;26:103) provide special challenges to the surgeon. When the tumor is in the upper rectum and an adequate margin can be obtained, a low anterior resection with primary anastomosis is performed. Preoperative chemotherapy and radiation is given (see following discussion of chemotherapy). Generally, all tumors outside of the reach of the examining finger can be resected this way. The circular end-to-end stapler revolutionized this operation by making the anastomosis easier to perform (Dis Colon Rectum 1999;42:1369). Pts who have a low anterior resection have a higher risk of anastomotic recurrence and are followed more intensively postoperatively. For those pts in whom a bulky tumor, local spread, or very distal location prevent an anterior approach, an abdominoperineal, or AP, resection of the rectum is performed, and a permanent sigmoid colostomy is created. This procedure is associated with significant sexual dysfunction (45%) and bladder dysfunction (30%). Sphinctersparing surgery (in the hands of specialists) can be considered if the tumor is 3 cm above the dentate line. Coloanal anastomosis has been described. Local excision may be reasonable for tumors that are within 8 cm of the dentate line, that are mobile, that have moderately or well-differentiated histology, and that are T1 or T2 by endoscopic US.

Metastatic disease to liver or lung can be approached surgically, especially if there is a solitary lesion. Adequate performance status, good hepatic reserve, and the ability to resect all evident disease are prerequisites. Tumor-free 5-yr survival is 20-30% for resection of hepatic mets, and the results are highly variable for lung mets (Gastroenterol Clin North Am 1997;26:103).

Chemotherapy: (GE 2008;134:1296) The utility of chemotherapy depends on the tumor stage and location. The goal of adjuvant rx is to treat micrometastases notidentified at operation (Semin Oncol 1999;26:545). Aspirin shows promise as an adjuvant Rx (Jama 2009;302:688). For stage III CRCs, the data currently support the use of 6 months of postoperative fluorouracil, leucovorin, and oxaliplatin. The role of adjuvant rx in stage II CRC is uncertain. Randomized trials don’t show benefit in this group that already has a good prognosis, but there could be a role for adjuvant rx in higher risk stage II cancers.

For stage II and III CRCs, preoperative combined radiation and fluorouracil-based chemotherapy has become the standard of care. For stage IV CRC, improvements in therapy have prolonged survival from 6 months to greater than 20 months. The newer agents, which include monoclonal antibodies to inhibit angiogenesis and growth factors, are very expensive.

Endoscopic therapy for obstruction: It is possible to treat acute obstruction with placement of expandable metal stents as a bridge to a single-stage resection or as palliation (Gastrointest Endosc 1998;47:277). Major complications are common, and the value of this approach needs additional evaluation.

The malignant polyp: See p 144.

Post-rx follow-up: About 50% of pts will have a recurrence of CRC after rx, and most of these will happen within 3 yr of surgery. Intensive programs of postoperative surveillance with monitoring of CBC, CMP, CEA, CXR, FOBT, frequent endoscopy, and abdominal CT scanning have been used in various combinations with disappointing results. The European Society for Medical Oncology (Ann Oncol 2008;19 Suppl 2:ii29) suggests options for a menu of more limited follow-up testing. A preoperative or perioperative colonoscopy should be done to clear the colon of synchronous lesions, repeated at 1 yr and be repeated every 3-5 yr thereafter. Periodic liver US or CT of chest and abdomen can be considered in pts at higher risk of recurrence. CXR has limited sensitivity. Postoperative CEA can be done q 3-6 months for 3 yr and q 6-12 months. Pts with elevations are evaluated fully for recurrence. About 30% of CRC does notproduce CEA, and up to 44% of pts with normal preoperative CEA have elevations with recurrence so CEA testing will have limited value (J Clin Oncol 1999;17:1312). Pts who are treated with a low anterior resection for rectal cancer have higher rates of local recurrence, and sigmoidoscopy or endoscopic US q 3-6 month for 2 yr can be considered (GE 2006;130:1865)