Machine Learning Models Cannot Replace Screening Colonoscopy for the Prediction of Advanced Colorectal Adenoma

Screening for colorectal cancer (CRC) continues to rely on colonoscopy and/or fecal occult blood testing since other (non-invasive) risk-stratification systems have not yet been implemented into European guidelines. In this study, we evaluate the potential of machine learning (ML) methods to predict advanced adenomas (AAs) in 5862 individuals participating in a screening program for colorectal cancer. Adenomas were diagnosed histologically with an AA being ≥ 1 cm in size or with high-grade dysplasia/villous features being present. Logistic regression (LR) and extreme gradient boosting (XGBoost) algorithms were evaluated for AA prediction. The mean age was 58.7 ± 9.7 years with 2811 males (48.0%), 1404 (24.0%) of whom suffered from obesity (BMI ≥ 30 kg/m²), 871 (14.9%) from diabetes, and 2095 (39.1%) from metabolic syndrome. An adenoma was detected in 1884 (32.1%), as well as AAs in 437 (7.5%). Modelling 36 laboratory parameters, eight clinical parameters, and data on eight food types/dietary patterns, moderate accuracy in predicting AAs with XGBoost and LR (AUC-ROC of 0.65–0.68) could be achieved. Limiting variables to established risk factors for AAs did not significantly improve performance. Moreover, subgroup analyses in subjects without genetic predispositions, in individuals aged 45–80 years, or in gender-specific analyses showed similar results. In conclusion, ML based on point-prevalence laboratory and clinical information does not accurately predict AAs.

Colorectal cancer screening and prevention: A brief review of the current guidelines and modalities

Colorectal cancer (CRC) remains a frequently addressed topic in primary care. Recent studies have been published detailing modifiable risk factors for CRC, as well as preventative measures. Providers must be up to date on screening recommendations and modalities. Colonoscopy is the preferred method of screening for CRC, and the screening recommendations in the United States were recently updated in 2020. It is also common for the practitioner to encounter patients who refuse colonoscopy but are willing to undergo alternative methods of testing. The COVID pandemic has also placed a burden on hospital resources, and colonoscopy may not be logistically feasible in some healthcare settings. Therefore, awareness of the guidelines for the various alternative modalities, along with their respective guidelines for frequency of screening is critical. This article provides a brief review of the risk factors associated with colon cancer, the screening modalities (including colonoscopy, sigmoidoscopy, CT colonography, fecal immunohistochemical testing (FIT), guaiac-based fecal occult blood testing (gFOBT), multi target stool DNA testing (MTs-DNA), and others) and the most recent screening recommendations for the general population.

The validity of self-reported cancer screening history and the role of social disadvantage in Ontario, Canada

Background Self-report may not be an accurate method of determining cervical, breast and colorectal cancer screening rates due to recall, acquiescence and social desirability biases, particularly for certain sociodemographic groups. Therefore, the aims of this study were to determine the validity of self-report of cancer screening in Ontario, Canada, both for people in the general population and for socially disadvantaged groups based on immigrant status, ethnicity, education, income, language ability, self-rated health, employment status, age category (for cervical cancer screening), and gender (for fecal occult blood testing). Methods We linked multiple data sources for this study, including the Canadian Community Health Survey and provincial-level health databases. Using administrative data as our gold standard, we calculated validity measures for self-report (i.e. sensitivity, specificity, positive and negative likelihood ratios, positive and negative predictive values), calculated report-to-record ratios, and conducted a multivariable regression analysis to determine which characteristics were independently associated with over-reporting of screening. Results Specificity was less than 70% overall and for all subgroups for cervical and breast cancer screening, and sensitivity was lower than 80% overall and for all subgroups for fecal occult blood testing FOBT. Report-to-record ratios were persistently significantly greater than 1 across all cancer screening types, highest for the FOBT group: 1.246 [1.189-1.306]. Regression analyses showed no consistent patterns, but sociodemographic characteristics were associated with over-reporting for each screening type. Conclusions We have found that in Ontario, as in other jurisdictions, there is a pervasive tendency for people to over-report their cancer screening histories. Sociodemographic status also appears to influence over-reporting. Public health practitioners and policymakers need to be aware of the limitations of self-report and adjust their methods and interpretations accordingly.

The validity of self-reported cancer screening history and the role of social disadvantage in Ontario, Canada

Background Self-report may not be an accurate method of determining cervical, breast and colorectal cancer screening rates due to recall, acquiescence and social desirability biases, particularly for certain sociodemographic groups. Therefore, the aims of this study were to determine the validity of self-report of cancer screening in Ontario, Canada, both for people in the general population and for socially disadvantaged groups based on immigrant status, ethnicity, education, income, language ability, self-rated health, employment status, age category (for cervical cancer screening), and gender (for fecal occult blood testing). Methods We linked multiple data sources for this study, including the Canadian Community Health Survey and provincial-level health databases. Using administrative data as our gold standard, we calculated validity measures for self-report (i.e. sensitivity, specificity, positive and negative likelihood ratios, positive and negative predictive values), calculated report-to-record ratios, and conducted a multivariable regression analysis to determine which characteristics were independently associated with over-reporting of screening. Results Specificity was less than 70% overall and for all subgroups for cervical and breast cancer screening, and sensitivity was lower than 80% overall and for all subgroups for fecal occult blood testing FOBT. Report-to-record ratios were persistently significantly greater than 1 across all cancer screening types, highest for the FOBT group: 1.246 [1.189-1.306]. Regression analyses showed no consistent patterns, but sociodemographic characteristics were associated with over-reporting for each screening type. Conclusions We have found that in Ontario, as in other jurisdictions, there is a pervasive tendency for people to over-report their cancer screening histories. Sociodemographic status also appears to influence over-reporting. Public health practitioners and policymakers need to be aware of the limitations of self-report and adjust their methods and interpretations accordingly.

Reddish Diarrhoea Caused by Excessive Ingestion of Watermelon: Mimicker of Bacterial Colitis

Bloody diarrhoea in children is indicative of serious diseases. Although bloody diarrhoea following bacterial colitis is well known, reddish diarrhoea caused by excessive ingestion of watermelon is unknown. A two year old girl who excessively ingested watermelon presented with repeated reddish diarrhoea. A kit for fecal occult blood testing revealed that the reddish diarrhoea did not contain blood. Reddish diarrhoea caused by excessive ingestion of watermelon can be a mimicker of bacterial colitis. The kit for fecal occult blood testing was useful for differential diagnosis. We should be aware that this pitfall can be hidden in daily nutrition.