New Clinical Model Predicts Need for Screening Colonoscopy
Douglas K. Rex, MD, MASGE, reviewing Imperiale TF, et al. Gut 2020 Sep 29.
A risk-stratified approach to colorectal cancer screening within the average-risk population is seldom used. In this approach, patients whose prescreening predicted prevalence of advanced neoplasia is high undergo colonoscopy, and those whose predicted prevalence is low undergo noninvasive screening.
In the new study, 50- to 80-year-olds undergoing first-time screening colonoscopy were used to derive a model that was based on age, gender, marital status (protective), college education (protective), smoking, significant ethanol use, nonsteroidal anti-inflammatory drug use, aspirin use, metabolic syndrome, red meat consumption, and physical activity score. The derivation set included 3025 participants, and the validation set included 1475 participants. The model was used to create brackets of low, intermediate, and high risk for advanced neoplasia. The low-risk, intermediate-risk, and high-risk groups comprised 23%, 59%, and 18%, respectively, of the study sample.
In the derivation set, the risk of advanced neoplasia in the low-, intermediate-, and high-risk groups was 1.5%, 7.06%, and 27.26%, respectively. In the validation set, the respective risks were 2.73%, 5.57%, and 25.79%.
COMMENTThe use of models to predict the risk of advanced neoplasia can improve the efficiency of colorectal cancer screening by sending patients with high risk and, perhaps, intermediate risk to colonoscopy and patients with low risk to noninvasive screening. The same group previously published a simpler model based on only 5 predictive factors. The current model is less practical because obtaining all the necessary information and computing scores would be more challenging in routine practice.
Note to readers: At the time we reviewed this paper, its publisher noted that it was not in final form and that subsequent changes might be made.
CITATION(S)
Imperiale TF, Monahan PO, Stump TE, Ransohoff DF. Derivation and validation of a predictive model for advanced colorectal neoplasia in asymptomatic adults. Gut 2020 Sep 29. (Epub ahead of print) (http://dx.doi.org/10.1136/gutjnl-2020-321698)
