Validation of Advanced Colorectal Neoplasm Risk Categories in a Prospective Cohort in Mexico

Overview

Worldwide, there are 1,361,000 new cases of colorectal cancers (CRC) annually, with 694,000 deaths. However, the incidence varies by up to a factor of 10x between high and low incidence countries (eg. USA vs Mexico, incidence rate of 42.54 vs 7.44 / 100,000 inhabitants). Mexico is considered a low-incidence country, with 8,651 new cases and 4,694 deaths annually. CRC is a preventable and detectable disease. Screening programs established in high-incidence countries have managed to reduce the incidence and mortality from this disease and it is considered a cost-effective strategy. In less developed countries where there are no screening programs for CRC, the highest number of deaths occurs despite having the lowest number of cases. It is recognized that a barrier to establishing a screening program in a country with low incidence and limited resources is cost-effectiveness. The prevalence of Advanced Colorectal Neoplasia (ACN) detected by screening colonoscopy in a Mexican cohort of 1172 INNSZ patients was 2.9%. In the US the prevalence is 7.6%. The number of colonoscopies to be performed to detect ACN was estimated at 34 for Mexico and 13 for the US, which suggests that the cost-effectiveness of screening colonoscopy could be 3 times lower in our country. In Mexico there is no national screening program for CRC. The eligible population (adults between 50 and 75 years old) for CRC screening is estimated in 20 million of Mexicans. It is recognized that Mexico does not have enough financial resources nor the infrastructure to screen the entire eligible population either by direct colonoscopy, or by FIT (fecal immunochemical test) followed by colonoscopy. With a 5% frequency of positive FIT, nearly 1,000,000 follow-up colonoscopies would be required annually in a population screening program. An alternative could be to offer screening based on risk, which means only offering screening to the highest-risk population. There are calculators to predict the risk of identifying ACN in a screening colonoscopy, however, none have been developed and validated in the Mexican population. The weight of the risk factors associated with ACN in the Mexican population could be different, so it is necessary to develop and validate an ACN risk calculator that allows the Mexican population to be stratified and to concentrate screening efforts on the population at highest risk.

This is a prospective cohort of subjects eligible for CRC screening in Mexico City. The main goal is to validate the APCS clinical score as a risk - stratification tool for screening colonoscopy in a Mexican population. The Asian-Pacific Colorectal Screening (APCS) score is a clinical tool used to stratify subjects according to the probability of identification of Advanced Colorectal Neoplasia (ACN) in a screening colonoscopy. ACN includes lesions larger than 10 mm with one or more of the following: a villous content, high-grade dysplasia, or carcinoma. The tool stratifies the population into three risk categories, considering: age, sex, smoking history, and first-degree family history of colorectal cancer. The total score obtained allows the stratification into average risk (0-1 point), moderate (2-3 points), and high risk (4-7 points). The prevalence of ACN in the original validation cohort of the APCS was 1.3% for average risk, 3.2% in moderate risk and 5.2% in high risk. Investigators previously evaluated the APCS score in a retrospective study of screening colonoscopy in Mexican population (n=1269). The prevalence of ACN was 2.6% vs 5.2% (p=0.027) for moderate risk and high-risk categories, respectively. To implement a risk-stratified screening program for the Mexican population, prospective validation was considered necessary. A prospective study would allow for collection and evaluation of additional variables that could potentially improve APCS score. FIT cut-off values have never been validated in Mexican population and are required to implement a national colorectal screening program. Prospective design will also allow for bio banking for future medical research. The following procedures will be performed: 1. Subjects will be invited to participate in the study, and if they accept, an informed consent will be given and explained to them. They will have a standardized interview dedicated to promoting CRC screening and complete a CRC risk factor questionnaire to determine their risk category. 2. Subjects may optionally participate in the collection of tissue, blood, and urine samples for future biomedical research. This procedure consists of taking a 15 ml sample of venous blood and 20 ml of urine. The tissue samples to be used will be those that have been taken for diagnostic purposes during the colonoscopy. Additional biopsies not required for medical care will not be taken. These samples will be used as long as they are no longer required to establish a diagnosis. 3. A Fecal Immunochemical Test (FIT) will be performed prior to the colonoscopy. Note: A positive FIT will have a Hb value \>20 ug/g of stool. 4. A colonoscopy will be performed. Any polyp identified will be resected (polypectomy) and analyzed histologically. NOTE: The subject and physician will be blinded to FIT test results and risk category. In subsequent medical consultation, the results of the risk score, the FIT test, and colonoscopy will be discussed. 5. If the patient does not perform any of the tests, they will be interviewed on a second occasion to complete a questionnaire of reasons of rejection. NOTE: It is anticipated that a proportion of subjects will verbally accept the screening, but the proposed tests will not be carried out (this will measure the percentage of acceptance of the recommendation). Given that age and gender are the most important risk factors, the population will be stratified simulating the Mexican population pyramid as follows: 50% male subjects and 50% female subjects. By age groups: 50-54 years (31%), 55-59 (24%), 60-64 (19%), 65-69 (14%), 70-75 (12%). The data collected from the participants will be entered into an electronic registry (Redcap) in real-time. A periodic review of the electronic registry will be carried out by personal of the protocol to ensure the quality of the information and that it is complete. The study does not contemplate external monitoring of the data. Statistical considerations * Sample size: for a difference in proportions of 2.6% (prevalence of ACN 5.2% vs. 2.6% in the high vs. moderate-risk population), 868 subjects per group are required to achieve a power of 80%, with a confidence level of 95 %, that is, 1736 subjects. It is anticipated that in 15% of the subjects the two screening tests will not be available, so a total sample of 2000 subjects is proposed. * Statistical proportions, means, and medians will be used to describe the population. The differences in proportions will be analyzed with the X2 test. Using a binary logistic regression model, variables associated with ACN risk will be identified. A p \< 0.05 will be considered statistically significant. The risk calculator will be developed using the variables identified in the univariate analysis and differential weight will be assigned to them according to risk. To determine the optimal cut-off point in FIT test (ng of hemoglobin per gram of feces): the investigators will use the distribution of the FIT score for patients with and without premalignant lesions and/or colorectal cancer to estimate sensitivity and specificity. according to different cut-off values. Subsequently, the investigators will use a decision model based on specificity/sensitivity ROC curves and according to different FIT cut-off points. The closer the ROC curve is to the upper left corner, the higher the predictive power of detecting colorectal lesions.