ABC of colorectal diseases. Colorectal neoplasia--III: Treatment and prevention

Images

Screening for large bowel neoplasms in individuals with a family history of colorectal cancer

Logistical problems associated with population screening for colorectal cancer are identified and the possibility of targeting screening to those with a familial predisposition to the disease is discussed. Evidence for a substantial genetic effect on the overall incidence of colorectal cancer is reviewed. The screening detection rate of colorectal neoplasms in relatives of patients with colorectal cancer has been shown to be higher than that expected in a non-selected population; the evidence that polypectomy will reduce future colorectal cancer risk in such individuals is explored. Recent advances in the molecular genetics of colorectal cancer susceptibility are reviewed; it is possible that a genetic test might be developed in the future which could identify at least a proportion of those at risk. Excluding financial considerations, the risk-benefit ratio of colonoscopy in a screened population is intimately related to the remaining risk of colorectal cancer in those who undergo the examination. At present, patients undergoing colonoscopy to investigate a positive faecal occult blood (FOB) test as part of a population-based screening programme include individuals with a familial predisposition as well as those without. About 20 per cent of all cases of colorectal cancer are associated with an obvious genetic predisposition, and the risk of cancer in their relatives is high. Because false positives occur with Haemoccult, the residual risk to the population who are FOB positive but do not have a familial trait may be sufficiently low that the dangers of colonoscopy could outweigh the potential benefits. Scotland has a high incidence of colorectal cancer, and analysis of recent Scottish incidence data shows an actuarial lifetime risk of developing this disease of one in 23 for men and one in 33 for women. As a family history of the disease increases that risk by two to four times and the neoplasms arise throughout the colon in such a group, there may be a case for offering colonoscopy to all first-degree relatives of those under 50 years of age at diagnosis, if not of all index cases of colorectal cancer.

Colonic fermentation and markers of colorectal-cancer risk

The aim of this study was to determine the effect of soluble fiber on indexes of colon-cancer risk in postpolypectomy and nonpolyp patients. Forty-five postpolypectomy and 49 nonpolyp volunteers completed 2-wk metabolic studies where half of the group received oat-bran supplements and the other half took wheat-brain supplements. Colonic biopsies taken before and after the intervention showed no difference in the index of thymidine colonic-crypt-cell labeling, thymidine-labeling pattern, or nuclear aberrations. Nevertheless, fecal pH was significantly reduced by 0.23 +/- 0.07 pH units (P less than 0.002) as an index of increased colonic fermentation on oat bran. This was not associated with increased basal breath hydrogen concentrations; fecal butyrate concentrations were higher on wheat bran. We conclude that soluble fiber as oat brain appears to have no advantage over wheat bran in modifying putative risk factors for colonic cancer.

Intermediate biomarkers of increased risk for colorectal cancer: comparison of different methods of analysis and modifications by chemopreventive interventions

Intermediate biomarkers of abnormal cell growth and development have recently been used in chemoprevention trials in attempts to identify the efficacy of chemopreventive agents in human subjects. Measurements carried out include those related to cell proliferation, differentiation, and gene structure and expression in the colon. Among modified patterns of cell proliferation identified by microautoradiographic or immunoperoxidase assays, a characteristic expansion in the size of the proliferative compartment has been observed in normal-appearing colorectal mucosa of human subjects with disease increasing cancer risk; the same patterns have been induced by chemical carcinogens in rodents. Moreover, this intermediate biomarker has been modulated by chemopreventive agents in both rodents and humans. Newer intermediate biomarkers being studied for application to human chemopreventive programs include normal and abnormal patterns of expression of mucins, intermediate filaments and cytoskeletal proteins, and the structure and expression of a variety of genes associated with normal and abnormal cell development. The application of these various intermediate biomarkers to chemoprevention studies is increasing the ability of investigators to analyze the effects of novel chemopreventive agents in the colon and in other organs.

Patient compliance with colorectal cancer screening in general practice

A randomized controlled trial to test patient compliance with screening for colorectal cancer in association with general practice health checks was carried out in six practices (three urban and three rural). A total of 1588 patients aged 45-64 years were randomized to one of four intervention groups. In the first group patients were posted a Haemoccult test (Kline Beckman) kit. This group was not invited for a health check. In the second group patients were posted the Haemoccult test kit, together with an invitation to attend for a health check. In the third group patients were posted an invitation for a health check, which explained that the patient would be offered the Haemoccult test kit by the nurse at the health check. In the fourth group patients were just invited for a health check. It was found that combining faecal occult blood testing with the health check did not reduce attendance at the health check--43.5% of patients attended when the Haemoccult test kit was offered by the nurse at the health check, 43.6% attended when a test kit was included with the invitation to attend the health check and 42.9% attended when the health check invitation was posted on its own. Overall, compliance with Haemoccult testing was not significantly increased by associating it with a health check (26.2% versus 25.5%) but compliance was higher when the faecal occult blood testing kit was enclosed with the health check invitation than when it was offered at the health check (31.7% versus 20.6%, P less than 0.001). It is easier and cheaper to combine various screening procedures. Although the overall use of the Haemoccult test in the study population was low, there is no reason why the relatively higher compliance rate obtained on posting the test kit with a health check invitation cannot be achieved in previously unscreened populations with higher expected compliance rates. However, faecal occult blood screening for colorectal cancer should not be undertaken on a population basis until its effectiveness in reducing mortality has been proven by randomized trial.

Cell proliferation in colorectal tumor progression: an immunohistochemical approach to intermediate biomarkers

Cell renewal in the large intestine mucosa is normally tied to a rigidly compartmentalized model. Immunohistochemical identification of cells in S phase through uptake of bromodeoxyuridine is the method of choice for detailed compartmental mapping of proliferation, while immunohistochemical detection of proliferation-associated antigens (Ki-67, PCNA, DNA polymerase alpha) provides information in advanced tumor cases. Mucosal hyperproliferation due to inflammation may be transient (self-limited colitis, Crohn's disease, acute radiation damage) or lasting (ulcerative colitis). Progressive shifting of the proliferation zone to the crypt surface (Stage II abnormality) is a late feature of irradiated rectal mucosa and subgroups of ulcerative colitis patients at high risk for cancer. Hyperproliferation and Stage II abnormality coexist in the mucosa of patients with colorectal neoplasia, but are mutually independent and correlated to different clinical and pathological features of the disease. These cytokinetic abnormalities are highly predictive markers of the adenoma-carcinoma sequence, but are not associated with de novo adenocarcinoma. Proliferation increases progressively in the subsequent steps of this sequence, except in early cancer.

[Role of flexible sigmoidoscopy in the early detection of colorectal tumors]

In a large endoscopic population screening study for colo-rectal polyps and cancer among 2,080 average-risk asymptomatic or with minor GI symptoms, adults of both sex, average age 56.2, were submitted to flexible sigmoidoscopy 60 cm as part of a preventive medical checkup. Procedure was very effective: 349 polyps were detected in 276 patients, with a total polyp incidence of 13.3%. In 174, more than 60% of such cases, endoscopic polypectomy was performed: 118 were tubular benign adenomas, 57 hyperplastic, 7 mixed types adenomas, 3 carcinomas in situ Dukes A, 2 villous adenomas one of them malignant, one leiomyoma and one rectal carcinoid. All in situ malignant polyps were removed endoscopically with excellent results. There was a high incidence of polyps (13.04%) in group aged 40-49, therefore changes on criteria for initiating routine flexible sigmoidoscopy must be considered. FOB tests (guaiac) showed a very low specific rate for diagnosis since only 24 out of 276 polyps bearing patients showed positive results, probably due to the fact that most polyps were non-ulcerating and small lesions. One third of the total polyps were out of the reach of the rigid scope and this procedure should be excluded in all screening programs of this type. Endoscopic polypectomy should be always an early procedure. In order to decrease the number of nondetected right side colonic lesions, non accessible to the 60 cm flexible scope, at time of polypectomy a simultaneous colonoscopy should be performed.(ABSTRACT TRUNCATED AT 250 WORDS)

The adenoma-adenocarcinoma sequence in the large bowel: variations on a theme

Most adenocarcinomas of the colorectum arise in a visible benign precursor lesion, the adenoma, which is a monoclonal proliferation of dysplastic nonmalignant epithelial cells. The resultant adenoma-adenocarcinoma sequence represents the predominant pathogenetic pathway, in contrast to de novo carcinoma. Therefore, the adenoma is a tempting endpoint for chemoprevention trials. The adenoma-adenocarcinoma sequence occurs in diverse clinical settings. In familial adenomatous polyposis (FAP) syndrome, autosomal dominant inheritance of the mutated APC (adenomatous polyposis coli) gene on chromosome 5q21 typically results in thousands of adenomas in the colorectum and in lesser numbers in the proximal small bowel. Adenocarcinoma usually develops in only a few of these adenomas, typically in the left colon and duodenum. In hereditary nonpolyposis colorectal cancer (HNPCC) syndrome, autosomal dominant inheritance of an unidentified gene appears to result in small numbers of adenomas which progress frequently to adenocarcinoma, predominantly in the right or transverse colon. In familial aggregation of colorectal cancer without a recognizable syndrome, cancer and/or adenomas occur in pedigree members. In "sporadic" cancers and adenomas, family history is absent and the tumors are mainly in the left colon. Colorectal adenomas have variable characteristics including size, shape (polypoid vs. flat), villous architecture, and dysplasia. A variety of oncogenes and tumor suppressor genes are altered during progression. Epigenetic factors are important as evidenced by the disappearance of adenomas in FAP patients after ileorectal anastomosis or treatment with the nonsteroidal antiinflammatory drug sulindac. Several variations on the theme of the adenoma-carcinoma sequence are evident. Identification of the inherited and acquired genetic alterations as well as the interacting environmental factors will provide a rational basis for chemoprevention.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of dietary calcium on the colonic luminal environment

Dietary supplementation with calcium may prevent the development of colorectal cancer. This mechanism may be related to fatty acid and bile salt chelation in the small bowel forming non-toxic calcium-soap compounds. Calcium may also act locally or systemically on the colonic mucosa. Faecal concentrations of free fatty acids and free bile acids were measured in 17 Sprague-Dawley rats (weighing 472 (39 g)) whose daily calcium intake had been trebled by enriching the chow and adding calcium lactate (24 g/l) to the drinking water. Mean (SEM) faecal concentrations of free bile acids were 33% less than in 19 controls (1.23 (0.15) v 1.82 (0.20) mg/g; p less than 0.001), whereas free fatty acid concentrations were 117% higher (14.68 (3.59) v 6.76 (2.41) mg/g; p less than 0.02). The 'direct' effect of calcium was assessed by organ culture of rat colonic explants in three different concentrations of calcium. Crypt cell production rate (measured by a stathmokinetic technique), which was (mean (SEM)) 4.80 (0.23) cells/crypt/h in control medium (Ca2+ = 2.14 mmol/l), fell by 43% when calcium concentration was doubled (p less than 0.05) and by a further 43% when the concentration was trebled (p less than 0.02). Calcium binds free fatty acids but not free bile acids intraluminally. Calcium has a direct antitropic action on colonic crypts.

[The status of colon cancer screening]

Recent research into an early detection of colorectal cancers shows a surprisingly great progress. In 1991, Mr. Hisamichi reports that a mass screening method for colorectal cancers through immunological fecal occult blood tests may result in fewer fatalities due to colorectal cancers. This paper describes the current situations and future prospects of UICC workshops in principle, as undermentioned in 1 through 6. 1. The target population should be clearly defined and selected in such a way that, given the likely sensitivity and specificity of the test, an acceptable predictive value will be achieved. 2. As far as possible, the target population should be limited to those at particular risk of the disease in order to give a reasonable prevalence of detection when such individuals at risk can be identified. 3. There should be a reasonable expectation of reaching the target population and of achieving reasonable compliance in response to invitations to attend. The expected level of compliance should be stated. 4. There should be a reasonable expectation that recommendations for further diagnostic evaluation will be complied with by individuals who have a positive test. 5. There should be an agreed-upon policy on the classification of borderline abnormalities and also on their management, as well as on recommendations for their follow-up. 6. There should be a reasonable expectation that recommendations for appropriate management of lesions discovered by a screening program will be complied with both by the individual with the lesion and by the physicians responsible for his health care.

Participation in mass screening for colorectal cancer: results of screening and rescreening from the Burgundy study

A controlled study was made in Burgundy (France) to assess the acceptability and efficiency of colorectal cancer screening using the Hemoccult test. A total of 91,000 people aged 45 to 74 years were included. The first screening campaign was made from January to July 1988 or 1989, and rescreening occurred from January to July 1990. During the first 4 months of the campaign, the General Practitioners (GPs) offered the test to their patients. It was then mailed to all subjects who had not completed it with a recall letter, if necessary, one month later. The overall compliance rate was 54.0% in the first campaign and 55.5% in the second campaign. A total of 63.7% subjects had completed at least one screening test over the two screening campaigns. Compliance was lower among men than women; compliance rate was also lower in the youngest and oldest age group than in the intermediate age groups. During the second screening campaign, the participation rate was 83.6% among participants in the first screening campaign and 20.9% among non-participants. Compliance during the medical offer phase was higher (81.4% during the first campaign, 82.9% during the second campaign) than during the postal offer phase (respectively, 33.8% and 28.5%). It was easier for GPs to propose the test when they had to offer it than when they had to prescribe it. These results suggest that a satisfactory participation rate can be obtained in France if GPs are actively involved in the screening programme. The final objective is to find a 25% difference in 5-year mortality by large bowel cancer between the screened and the test population.

Prevention and management of recurrent rectal cancer

Increased attention to surgical indications and technique can minimize the risks of recurrent rectal cancer. In addition, the use of adjuvant chemoradiation therapy has been shown to further decrease the risks of recurrent rectal cancer. In the event a recurrence develops, surgical therapy, combined with radiation therapy, can result in local control in as many as 75% of patients and long term survival in 25% to 40% of patients.

The cost of screening for colorectal cancer

STUDY OBJECTIVE

The aim was to make projections of the likely costs and yield resulting from the implementation of a faecal occult blood screening programme for colorectal cancer.

DESIGN

Cost and clinical data were derived from the MRC colorectal screening trial currently in progress in Nottingham, UK.

SETTING

The above data were used as the basis for modelling the likely implications were the trial to be reproduced as a screening programme within a "typical" family practitioner committee area.

MAIN RESULTS

For an average family practitioner committee area with a target population of 75,000 subjects aged 50-74 years, the initial screening round might be expected to detect 85 cancers at a total cost of approximately 250,000 pounds. This represents a cost per cancer detected of 2700 pounds and a cost per person screened of approximately 5 pounds. For subsequent screening rounds, total costs might be expected to fall although average costs are likely to remain approximately constant.

CONCLUSIONS

The model is successful in generating "order of magnitude" estimates for the costs of implementation of a screening programme for colorectal cancer. As benefit estimates are not yet available, however, no cost-effectiveness analysis can be undertaken at this stage. In general, sensitivity analyses reveal that programme costs are more sensitive to changes in clinical variables, especially detection and compliance rates, than they are to variations in the costs of resource inputs. A screening programme with a more elaborate protocol than that currently employed in the Nottingham trial will entail considerable cost increases.

Cost-effectiveness analysis of screening for colorectal cancer in Japan

To clarify the best cost-effective screening strategy for colorectal cancer in Japan, the cost-effectiveness ratio was compared among six currently performed procedures. The analysis was made using a simulation model to estimate long-term cost and effectiveness of the screening programs. In the screening by the immunological fecal occult blood test (FOBT), a comparison between the one- and two-day fecal collection methods indicated that the latter was more cost-effective than the former. A comparison was also made on the four workup methods: barium enema (BE) alone, a combination of BE and sigmoidoscopy (BE + SIG), total colonoscopy (TCF) alone, and a combination of BE and TCF (BE + TCF). The cost-effectiveness ratio was the lowest in the method using TCF alone, followed by those based on BE alone and BE + TCF, and the highest in the BE + SIG method. The superiority of TCF alone strategy was stable over a range of estimates such as the sensitivity of diagnostic tests, the probability of complications due to TCF, etc. It is concluded that a combination of the two-day FOBT and TCF yields the best cost-effectiveness.

Cost-effectiveness analysis of screening for gastric cancer in Japan

Cost-effectiveness analysis of gastric cancer screening in Japan was performed, and the cost-effectiveness ratio was compared with that of colorectal cancer screening. The analytical model was the same as that adopted in our previous study on colorectal cancer screening (Tsuji et al. 1991). The results indicated that the cost per case-fatality prevented by colorectal cancer screening was 5.5 and 2.7 times more expensive for males and females, respectively, than that by gastric cancer screening. The age of the population influenced the cost-effectiveness of screening programs. The cost-effectiveness ratio in gastric cancer screening became negative among males older than the age of 65 years, suggesting that the total cost for prevention and control of deaths by gastric cancer in these age populations is saved by the screening program.

[The screening model for early diagnosis of colorectal cancer in general population]

In this study, 75,813 asymptomatic individuals above 30 in Jiashan County of Zhejiang Province were recruited in the screening program. Based on results of risk assessment using a mathematical model (expressed as AD value) and Reverse Passive Hemagglutination Fecal Occult Blood test (RPHA FOB). 62,667 subjects (82.7%) who completed two procedures were divided into 4 groups: 1) AD greater than or equal to 0.3 FOB (+) accounts for 1.2% of total; 2) AD less than 0.3 FOB (+) 3.0%; 3) AD greater than or equal to 0.3 FOB (-) 26.8%; 4) AD less than 0.3 FOB (-) 68.9%. Individuals in groups 1 and 2 were subjected to 60 cm fiberoptic colonoscopic examination; and those of group 3 with AD greater than or equal to 0.5 or with AD less than 0.5 but a positive repeated FOB test was also eligible for 60 cm colonoscopy. Out of 4,299 eligible subjects 3,162 were examined with 60 cm sigmoidoscopy (73.6%). 21 cases of colorectal cancer and 331 cases of polyps (including 75 adenomas) were detected. Among 21 cancers known of Dukes' stages 10 were at early stages (10 Dukes' A and 5 Dukes' B) accounting for 71.5%. According to the screening protocol individuals with negative 60 cm colonoscopic findings and consistently positive FOB results indicate that the lesion may be located above the spleen flexure and require a full fiberoptic colonoscopic examination (150 cm) and/or x-ray air-barium enema.

Folate intake and carcinogenesis of the colon and rectum

Experimental evidence suggests that folate depletion plays a role in carcinogenesis. A case-control study examining folate intake was conducted. Some 428 colon and 372 rectal cancer cases with matched neighbourhood controls were interviewed regarding usual intake of foods, including food preparation. Unadjusted folate was not associated with risk of either cancer. Controlling for kilocalories, odds ratios (ORs) for those with the highest folate intake were 0.5 (95% confidence interval (CI): 0.24-1.03) and 0.31 (95% CI: 0.16-0.59) for females and males for rectal cancer. There was no change in colon cancer risk associated with folate intake. There was an indication of an interaction of folate and alcohol intake; the difference in risk associated with low and high folate intake was highest for males in the highest alcohol category. Associations were of similar magnitude for other dietary factors correlated with folate. It appears that intake of folate or a correlated factor may be negatively related to risk of rectal cancer.

Failure of CA19-9 to detect asymptomatic colorectal carcinoma

Serum CA19-9 levels have been measured in 34 patients with asymptomatic colorectal cancer, 39 age and sex matched subjects with healthy colons (as assessed at full colonoscopy) and 55 patients known to have liver metastases from primary colorectal cancers. In subjects with asymptomatic cancer the median CA19-9 was 12.75 U ml-1 (0.0-280.7 U ml-1), in the healthy controls the median CA19-9 was 12.80 U ml-1 (0.0-88.9 U ml-1) and in those with liver metastases was 62.5 U ml-1 (4.8-458.0 U ml-1). Levels were significantly higher in patients with metastatic disease than in patients with asymptomatic tumours or the healthy controls, however there was no significant difference between the asymptomatic group and the controls. Using an upper limit of normal of 37 U ml-1, the sensitivity of CA19-9 was 60.3% for the detection of colorectal cancer with liver metastases but only 17.6% for asymptomatic cancer. Serum CA19-9 estimation is of no value as a means of screening for asymptomatic colorectal cancer.