D. James B. St. John

Cancer risk in relatives of patients with common colorectal cancer.

Screening for colorectal cancer continues to be controversial [1, 2]. The role of fecal occult blood tests is being evaluated [3-8], but definitive results are unlikely before 1995. One intervening strategy is to improve screening for people at high risk, for example, those with a family history of colorectal cancer. Secondary prevention of colorectal cancer (detection and resection of adenomas and early stage cancers to prevent fatal cancer) is already well accepted in familial adenomatous polyposis and hereditary nonpolyposis colorectal cancer (HNPCC) [9] and is somewhat accepted in ulcerative colitis [10-12]. Close relatives of patients with common (sporadic) colorectal cancer (those cancers not associated with familial adenomatous polyposis, HNPCC, or chronic inflammatory bowel disease) may have an increased risk for colorectal cancer [2], and selective screening has been suggested for these relatives [13-16]. Most of the evidence for increased risk is based on cancer mortality studies [17-19] and cancer incidence studies [20-27]. However, problems with study design have led to conflicting estimates of risk ranging from no increase [21] to 6.3-fold [22] and 8-fold [20] increases in risk for first-degree relatives. Pedigree analysis has suggested that an inherited susceptibility may be present in common colorectal cancer [28, 29], but the yield of neoplastic lesions in screened relatives has been low in most [30-32] but not all prevalence studies [15, 33]. Accordingly, we did a casecontrol family study to assess risk for colorectal cancer in relatives and to define possible markers for any increased risk. Methods The study used a matched casecontrol design [34] to compare the occurrence of cancer in relatives between patients with common colorectal cancer (case patients) and patients without cancer (control patients). Case and control patients were selected from one surgeons practice and were matched by gender, age (within 5 years and usually within 2 years), and country of birth or upbringing. Patients Case patients were selected from a series of more than 2400 patients with colorectal cancer treated between 1952 and 1985. All 576 patients alive at the time of the survey (June 1985 to May 1986) and residing in Australia were reviewed for eligibility. Detailed clinical, surgical, and pathologic information was available for every patient [35, 36]. Exclusion criteria were 1) colorectal cancer associated with ulcerative colitis, Crohn disease, familial adenomatous polyposis, or recognized HNPCC [37]; and 2) referral to the surgeon because he previously had treated a relative or spouse for colorectal cancer or adenoma. The minimum criteria for diagnosis of the HNPCC were three affected first-degree relatives with at least two relatives having colorectal cancer diagnosed before 50 years of age, or multiple primary cancers, or colorectal cancer plus another syndrome cancer [38]. Control patients were selected from patients who had surgery for nonmalignant disease during the same period and had no known history of neoplastic disease. Data Collection Questionnaires were mailed concurrently to case patients and their matched control patients. Accompanying letters made case and control patients equally aware that the study was about family history of colorectal cancer. The questionnaires covered personal information; information about first-degree relatives (parents, siblings, and children surviving to the age of 12 years), including year of birth, occurrence of colorectal cancer, age at diagnosis, occurrence of other types of cancer, age at death, and cause(s) of death when relevant; and information about second-degree relatives and spouses. Each case patient and control was contacted by telephone to review responses and to assess the degree of certainty about listed diagnoses. Care was taken to question case and control patients to the same degree. In many instances, responses were reviewed during a subsequent visit. Close relatives were encouraged to help collect information. All questionnaires were masked to prevent identification of their casecontrol status and were then assessed by an independent gastroenterologist to classify each relative as evaluable or nonevaluable according to whether information about his or her medical history was available up to the time of death or, in the case of living relatives, up to the time of completion of the questionnaire. Half-brothers and half-sisters were excluded. Many subjects were contacted again at this stage. Verification of diagnoses was sought for all reported cases of bowel cancer, other abdominal cancer, and abdominal surgery of unspecified type using hospital or physician records, Australian state cancer registries, and death registries. Colorectal cancers in relatives were classified as definite when confirmation of the diagnosis was obtained from a medical source; as probable when relatives were able to provide convincing information about the diagnosis (such as name of the surgeon and hospital as well as details about the surgical procedure) but where medical confirmation could not be obtained; and as possible when relatives were stated to have had colorectal cancer but when the evidence was less convincing. Final Selection of Patients for the Study Of the 576 patients with colorectal cancer alive at the time of the survey, 20 were excluded because cancer was associated with ulcerative colitis, Crohn disease, or familial adenomatous polyposis and 22 were excluded because of failure to complete questionnaires (15 had terminal illness, 5 refused, and 2 could not be contacted). Five potential control patients failed to complete questionnaires, and suitable replacements were recruited. Eleven case patients (and their respective control patients) were classified as ineligible and were excluded from the study; two proved to be members of families with recognized HNPCC, and the other nine had been referred to the surgeon because he previously had treated a relative or spouse for colorectal cancer. The research protocol was approved by the Ethics Review Advisory Committee at Alfred Hospital, Melbourne, Australia. Statistical Analysis Case-Control Analysis The data were first analyzed as a casecontrol study, with exposure defined by the presence of disease among (subsets of) relatives. The proportion of case patients with affected relatives was compared with the proportion of control patients, with the odds ratio and standard error calculated by standard methods using logistic regression [39]. Risk to Relatives of a Case Patient Compared with Relatives of a Control Patient A cohort analysis of disease in relatives, which takes into account the matching of casecontrol pairs and the number of relatives and allows for relatives to be dependent in their colorectal cancer status (as is the case if true familial aggregation exists for colorectal cancer), was also done [40]. For each case patient, i, let there be ni relatives of a predetermined type (for example, mother, parent, or sibling), of whom xi also have colorectal cancer. Let the matched control patient, i, have mi such relatives of whom yi have colorectal cancer. Let ti = xi + yi and Ni = ni + mi and the true odds ratio, ORi, for cancer in a relative of a case patient compared with cancer in a relative of a control patient be defined as (E[xi]E[mi yi])/(E[yi]E[ni xi]). Suppose a true common odds ratio exists; for example, ORi = OR for all i. The Mantel-Haenszel estimator is ORMH = Ri/Si, where Ri = xi(mi yi)/Ni and Si = yi(ni xi)/Ni. It is asymptotically normal, even if nonindependence exists within sets of relatives. A consistent and robust test statistic for Ho: OR = 1 is EQUATION OMITTED which has an approximate chi-square distribution with one degree of freedom, independent of the underlying true distribution for the {xi} and {yi}. Because log(ORMH) has more nearly a normal distribution than ORMH, an approximate test-based standard error, SE, for log(ORMH) was calculated by assuming (log[ORMH]/SE)2 = T. An approximate asymmetric 95% CI for ORMH is (elog[ORMH sup] 1.96SE, elog[ORMH sup] + 1.96SE). The same analysis was applied to subsets of casecontrol pairs selected according to a predetermined characteristic of the case patients (for example, age of onset less than 45 years). Analysis of Cumulative Incidence among Relatives of (Subsets of) Case and Control Patients Parents, siblings, and children were pooled. A Kaplan-Meier survival curve, S(t), was calculated for time free of colorectal cancer, t, based on age at diagnosis in relatives [41]. (Age at death was treated as a censoring time.) Cumulative incidence to age, a, was then defined as 1 S(a) and was plotted for 10-year increments in age. Comparison of survival, and therefore of cumulative incidence, between groups was evaluated by the two-tailed generalized Wilcoxon test according to Gehan [42]. Results Characteristics of Case and Control Patients The study group comprised 523 cases (250 men, 273 women; mean age, 71.2 years [SD, 10.3 years]), and 523 gender-matched control patients (mean age, 69.9 years [SD, 10.2 years]). Of the case patients, 426 (81%) were born in Australia. Thirteen (2.5%) case patients had been treated for synchronous cancers of the large bowel and 30 (5.7%) for metachronous cancers. The primary lesions were in the proximal colon (proximal to the splenic flexure) in 83 case patients, in the distal colon in 148, and in the rectum or at the rectosigmoid junction in 338. Colorectal cancer was diagnosed before age 45 years in 82 (16%) case patients, between ages 45 and 54 years in 147 (28%) patients, between ages 55 and 64 years in 173 (33%), and at 65 years in 121 (23%). A diagnosis of colorectal cancer was made at least 5 years before the survey in 436 (83%) case patients. Characteristics of Relatives The case and control patients reported a total of 7530 first-degree relatives, of whom all but 37 were evaluable. This left 3857 relati

Use of Molecular Tumor Characteristics to Prioritize Mismatch Repair Gene Testing in Early-Onset Colorectal Cancer

PURPOSE The relationships between mismatch repair (MMR) protein expression, microsatellite instability (MSI), family history, and germline MMR gene mutation status have not been studied on a population basis. METHODS We studied 131 unselected patients with colorectal cancer diagnosed younger than age 45 years. For the 105 available tumors, MLH1, MSH2, MSH6, and PMS2 protein expression using immunohistochemistry (IHC) and MSI were measured. Germline DNA was screened for hMLH1, hMSH2, hMSH6, and hPMS2 mutations for the following patients: all from families fulfilling the Amsterdam Criteria for hereditary nonpolyposis colorectal cancer (HNPCC); all with tumors that were high MSI, low MSI, or that lacked expression of any MMR protein; and a random sample of 23 with MS-stable tumors expressing all MMR proteins. RESULTS Germline mutations were found in 18 patients (nine hMLH1, four hMSH2, four hMSH6, and one hPMS2); all tumors exhibited loss of MMR protein expression, all but one were high MSI or low MSI, and nine were from a family fulfilling Amsterdam Criteria. Sensitivities of IHC testing, MSI (high or low), and Amsterdam Criteria for MMR gene mutation were 100%, 94%, and 50%, respectively. Corresponding positive predictive values were 69%, 50%, and 75%. CONCLUSIONS Tumor IHC analysis of four MMR proteins and MSI testing provide a highly sensitive strategy for identifying MMR gene mutation-carrying, early-onset colorectal cancer patients, half of whom would have been missed using Amsterdam Criteria alone. Tumor-based approaches for triaging early-onset colorectal cancer patients for MMR gene mutation testing, irrespective of family history, appear to be an efficient screening strategy for HNPCC.

Evaluation of new occult blood tests for detection of colorectal neoplasia

BACKGROUND Hemoccult II, the guaiac-based fecal occult blood test used in most colorectal cancer screening programs, has an unsatisfactory sensitivity for asymptomatic colorectal neoplasms. We evaluated the relative performance of four fecal occult blood tests, directed against various components of the hemoglobin molecule. METHODS All tests, Hemoccult II, HemoccultSENSA (a more sensitive guaiac test), HemeSelect (an immunochemical test specific for human hemoglobin), and HemoQuant (the heme-porphyrin assay), were performed by 107 patients with symptomatic colorectal cancer and 81 patients with predominantly asymptomatic adenoma. Hemoccult-SENSA and HemeSelect were performed by 1,355 screenees. RESULTS HemeSelect and Hemoccult-SENSA had significantly higher sensitivity for colorectal cancer (97% and 94%, respectively) than the other tests. HemeSelect had the highest sensitivity for adenomas; in 45 patients with large (> or = 10 mm) adenomas, sensitivity was 76% for HemeSelect, 60% for HemoccultSENSA, and 42% for both Hemoccult and HemoQuant. In the screenees, estimated specificity was 97.8% for HemeSelect and 96.1% for Hemoccult-SENSA. CONCLUSIONS HemeSelect and Hemoccult-SENSA have the highest levels of sensitivity for detection of colorectal neoplasia, but the immunochemical test HemeSelect provides the best combination of specificity and sensitivity.

Choice of fecal occult blood tests for colorectal cancer screening: recommendations based on performance characteristics in population studies: a WHO (World Health Organization) and OMED (World Organization for Digestive Endoscopy) report

OBJECTIVE: There is now strong evidence that screening for colorectal cancer with fecal occult blood tests (FOBTs) is effective in reducing the incidence and mortality of this disease. Various FOBTs are now available with a wide range of evidence supporting their use. The purpose of this study was to review published data on the performance of these FOBTs to provide recommendations for their effective use in screening. METHODS: A joint committee representing the World Health Organization and the World Organization for Digestive Endoscopy was established for this study. A process was designed that would search the literature systematically for evidence of FOBT performance. Criteria for including studies in this paper were established based on study design, cohort size, and performance variables reported. RESULTS: Of the guaiac tests, Hemoccult SENSA had the highest sensitivity for cancer and adenomas but a high test positivity. It had a better readability than the older Hemoccult II test. Immunochemical tests, HemeSelect, FlexSure OBT, and Immudia Hem Sp have acceptable performance characteristics and are easier for participants to use but are more expensive. These tests have been well studied in large cohorts, but only Immudia Hem Sp is commercially available. CONCLUSIONS: At present, there is no extensively studied FOBT that fulfills the needs for all target populations worldwide. Choice of FOBT should take into account population dietary compliance and colonoscopy resources: The more sensitive newer tests should be used if dietary compliance is good (in the case of guaiac tests) and colonoscopy resources are adequate for diagnostic workup of people who test positive. Immunochemical tests remove the difficulties created by diet and drug restrictions and are more amenable to standardized development and quality control.

Characteristics of small bowel carcinoma in hereditary nonpolyposis colorectal carcinoma

Small bowel carcinoma is uncommon. However, hereditary nonpolyposis colorectal carcinoma (HNPCC) patients are at increased risk of small bowel carcinoma. The purpose of this study was to characterize small bowel tumors in HNPCC patients.

Optimal Dietary Conditions for Hemoccult Testing

Preliminary rehydration of Hemoccult II slides increases slide sensitivity for blood and lowers the false-negative rate for colorectal cancer. To test the effect of this modification on the false-positive rate due to peroxidase-containing foods, 156 healthy young subjects crossed between diets of differing peroxidase content. Hemoccult II slides were prepared in duplicate for testing with and without rehydration. When developed without rehydration, only seven (0.4%) of 1856 slides were positive, the positive results occurring with challenge diets containing 250 g rare red meat. With preliminary rehydration, 53 (5.7%) of 926 slides were positive, in 26 (17%) of the 156 subjects on a challenge diet that included rare red meat and uncooked fruit and vegetables. Well-cooked red meat also gave positive tests, Diets excluding red meat but with large quantities of uncooked fruit and vegetables produced only five positive results in 314 rehydrated tests, in 3 of 53 subjects. On a strict low-peroxidase diet, two of 310 tests, in 2 of 52 subjects, were positive. Rehydration of Hemoccult slides and exclusion of red meat and certain other high-peroxidase foods should give optimal sensitivity and specificity for colorectal cancer detection. The effect of these conditions on yield of lesions and false-positive rates should be tested in screening programs.

Choice of fecal occult blood tests for colorectal cancer screening: recommendations based on performance characteristics in population studies

OBJECTIVE:There is now strong evidence that screening for colorectal cancer with fecal occult blood tests (FOBTs) is effective in reducing the incidence and mortality of this disease. Various FOBTs are now available with a wide range of evidence supporting their use. The purpose of this study was to review published data on the performance of these FOBTs to provide recommendations for their effective use in screening.METHODS:A joint committee representing the World Health Organization and the World Organization for Digestive Endoscopy was established for this study. A process was designed that would search the literature systematically for evidence of FOBT performance. Criteria for including studies in this paper were established based on study design, cohort size, and performance variables reported.RESULTS:Of the guaiac tests, Hemoccult SENSA had the highest sensitivity for cancer and adenomas but a high test positivity. It had a better readability than the older Hemoccult II test. Immunochemical tests, HemeSelect, FlexSure OBT, and Immudia Hem Sp have acceptable performance characteristics and are easier for participants to use but are more expensive. These tests have been well studied in large cohorts, but only Immudia Hem Sp is commercially available.CONCLUSIONS:At present, there is no extensively studied FOBT that fulfills the needs for all target populations worldwide. Choice of FOBT should take into account population dietary compliance and colonoscopy resources: The more sensitive newer tests should be used if dietary compliance is good (in the case of guaiac tests) and colonoscopy resources are adequate for diagnostic workup of people who test positive. Immunochemical tests remove the difficulties created by diet and drug restrictions and are more amenable to standardized development and quality control.

Screening and preventive behaviors one year after predictive genetic testing for hereditary nonpolyposis colorectal carcinoma

Prevention benefits from predictive genetic testing for cancer will only be fully realized if appropriate screening is adopted after testing. The current study assessed screening and preventive behaviors during 12 months after predictive genetic testing for hereditary nonpolyposis colorectal carcinoma (HNPCC) in an Australian clinical cohort.

After hMSH2 and hMLH1—what next? Analysis of three-generational, population-based, early-onset colorectal cancer families

The aim of our study was to examine the role of genetic factors on early‐onset colorectal cancer after excluding the impact of germline mutations in the two major mismatch repair genes. A total of 131 incident probands, under 45 years at diagnosis of a first primary colorectal cancer selected from the Victorian Cancer Registry, and their first‐and second‐degree relatives, were interviewed. Germline DNA from all 12 probands with a family history meeting the modified Amsterdam Criteria for Hereditary Non‐Polyposis Colorectal Cancer (HNPCC) and a random sample of 31 of the remaining probands was screened for mutations in hMSH2 and hMLH1 via manual sequencing. Germline mutations were identified in 6 of the 131 probands (5%), all from the “HNPCC” families. Of the remaining 125 probands, 51 (41%) reported at least one first‐or second‐degree relative with colorectal cancer with an excess of colorectal cancer in first‐degree relatives (SMR = 2.7, 95% CI = 1.7–4.1, p < 0.001). The lifetime risk to age 70 for first‐degree relatives was 8.0% (5.0–12.8%), compared to the Victorian population risk of 3.2% (p = 0.01). The best fitting major gene model was a recessively‐inherited risk of 98% to age 70 (95% CI = 24–100%) carried by 0.17% of the population and would explain 15% of all colorectal cancer in cases with a diagnosis before age 45. Early‐onset colorectal cancer is strongly familial even after excluding families found to be segregating a mutation in either of the 2 major mismatch repair genes. There is evidence for a role of yet to be identified genes associated with a high recessively‐inherited risk of colorectal cancer.

Comparison of the Specificity and Sensitivity of Hemoccult and HemoQuant in Screening for Colorectal Neoplasia

OBJECTIVE To compare the Hemoccult II and HemoQuant tests regarding their specificity and sensitivity in screening for colorectal neoplasia. DESIGN Cross-sectional study in which subjects underwent the two tests in parallel, after excluding dietary hemes and peroxidase-rich foods. HemoQuant results were analyzed for three different upper limits of normality (1.5, 2.0, 3.0 mg/g feces). SETTING A university hospital and the surrounding community. PARTICIPANTS A total of 150 healthy volunteers, 124 patients with colorectal cancer, and 86 patients with adenoma. MAIN OUTCOME MEASURES Blinded comparison of the specificity and the sensitivity of the two tests for colorectal cancer and adenoma. RESULTS Test specificity was 99.3% with Hemoccult and was 92.7%, 94.7%, and 97.3% with HemoQuant, depending on the cutoff point; differences between Hemoccult and HemoQuant were significant when cutoffs of 1.5 and 2.0 mg/g were used in HemoQuant testing (6.6% [95% CI, 2.3 to 11.1] and 4.7% [CI, 0.8 to 8.5], respectively). Test sensitivity for colorectal cancer at all sites was 89.5% with Hemoccult and was 83.1%, 74.2% and 62.9% with HemoQuant, for the 1.5, 2.0, and 3.0 mg/g cutoffs, respectively; differences were significant with the 2.0 and 3.0 mg/g cutoff points (6.4% [CI, 6.7 to 24.0] and 26.6% [CI, 17.4 to 35.9%], respectively). The two tests had similar levels of sensitivity for cancers proximal to the splenic flexure, but sensitivity was substantially lower with HemoQuant for the more distal cancers. For all adenomas, test sensitivity was 30.2% with Hemoccult and ranged from 45.4% to 22.1% with HemoQuant. CONCLUSIONS Although HemoQuant provides a precise measurement of fecal heme and its porphyrin degradation products, the tests performance characteristics in the detection of colorectal neoplasia are less satisfactory than those of Hemoccult II, a qualitative test for the presence of heme.