Post-polypectomy colonoscopy surveillance: Can we improve the diagnostic yield?

Machine learning in colorectal polyp surveillance: A paradigm shift in post-endoscopic mucosal resection follow-up

Colorectal cancer remains a major health concern, with colorectal polyps as key precursors. Endoscopic mucosal resection (EMR) is a common treatment, but recurrence rates remain high. Traditional surveillance strategies rely on polyp characteristics and completeness of the resection potentially missing key risk factors. Machine learning (ML) offers a transformative approach by integrating patient-specific data to refine risk stratification. Recent studies highlight ML models, such as Extreme Gradient Boosting, which outperform conventional methods in predicting polyp recurrence within one-year post-EMR. These models incorporate factors like age, smoking status, family history, and pathology, optimizing follow-up recommendations and minimizing unnecessary procedures. Artificial intelligence (AI)-driven tools and web-based calculators enhance clinical workflow by providing real-time, personalized risk assessments. However, challenges remain in external validation, model interpretability, and clinical integration. Future surveillance strategies should combine expert judgment with AI insights to optimize patient outcomes. As gastroenterology embraces AI, ML-driven surveillance represents a paradigm shift, advancing precision medicine in colorectal polyp management. This editorial explores AI’s role in transforming post-EMR follow-up, addressing benefits, limitations, and future directions.

Comparative Study of Predictive Models for the Detection of Patients at High Risk of Inadequate Colonic Cleansing

Background: Various predictive models have been published to identify outpatients with inadequate colonic cleansing who may benefit from intensified preparations to improve colonoscopy quality. The main objective of this study was to compare the accuracy of three predictive models for identifying poor bowel preparation in outpatients undergoing colonoscopy. Methods: This cross-sectional study included patients scheduled for outpatient colonoscopy over a 3-month period. We evaluated and compared three predictive models (Models 1-3). The quality of colonic cleansing was assessed using the Boston Bowel Preparation Scale. We calculated the area under the curve (AUC) and the corresponding 95% confidence interval for each model. Additionally, we performed simple and multiple logistic regression analyses to identify variables associated with inadequate colonic cleansing and developed a new model. Results: A total of 649 consecutive patients were included in the study, of whom 84.3% had adequate colonic cleansing quality. The AUCs of Model 1 (AUC = 0.67, 95% CI [0.63-0.70]) and Model 2 (AUC = 0.62, 95% CI [0.58-0.66]) were significantly higher than that of Model 3 (AUC = 0.54, 95% CI [0.50-0.58]; p < 0.001). Moreover, Model 1 outperformed Model 2 (p = 0.013). However, the new model did not demonstrate improved accuracy compared to the older models (AUC = 0.671). Conclusions: Among the three compared models, Model 1 showed the highest accuracy for predicting poor bowel preparation in outpatients undergoing colonoscopy and could be useful in clinical practice to decrease the percentage of inadequately prepared patients.

Using New Hampshire Colonoscopy Registry data to assess United States and European post-polypectomy surveillance guidelines

BACKGROUND

Our goal was to compare the updated European Society of Gastrointestinal Endoscopy (ESGE) and United States Multi-Society Task Force on Colorectal Cancer (USMSTF) high risk groups in predicting metachronous advanced neoplasia on first follow-up colonoscopy and long-term colorectal cancer (CRC).

METHODS

We compared advanced metachronous neoplasia risk (serrated polyps ≥ 1 cm or with dysplasia, advanced adenomas [≥ 1 cm, villous, high grade dysplasia], CRC) on first surveillance colonoscopy in patients with high risk findings according to ESGE versus USMSTF guidelines. We also compared the positive and negative predictive values (PPV, NPV) of both guidelines for metachronous neoplasia.

RESULTS

The risk for metachronous neoplasia in our sample (n = 20 458) was higher in the high risk USMSTF (3 year) (13.6 %; 95 %CI 12.3-14.9) and ESGE groups (13.6 %; 95 %CI 12.3-15.0) compared with the lowest risk USMSTF (5.1 %; 95 %CI 4.7-5.5; P < 0.001) and ESGE categories (6.3 %; 95 %CI 6.0-6.7; P < 0.001), respectively. Adding other groups such as USMSTF 5-10-year and 3-5-year groups to the 3-year category resulted in minimal change in the PPV and NPV for metachronous advanced neoplasia. High risk ESGE (hazard ratio [HR] 3.03, 95 %CI 1.97-4.65) and USMSTF (HR 3.07, 95 %CI 2.03-4.66) designations were associated with similar long-term CRC risk (CRC per 100 000 person-years: USMSTF 3-year group 3.54, 95 %CI 2.68-4.68; ESGE high risk group: 3.43, 95 %CI 2.57-4.59).

CONCLUSION

Performance characteristics for the ESGE and USMSTF recommendations are similar in predicting metachronous advanced neoplasia and long-term CRC. The addition of risk groups, such as the USMSTF 5-10-year and 3-5-year groups to the USMSTF 3-year category did not alter the PPV or NPV significantly.