Risk factor analysis of malignant adenomas detected during colonoscopy

Abstract

Background: Several studies have shown that colorectal adenomas are the most important precancerous lesions. The colonoscopic identification of groups with the high risk of malignant colorectal adenomas remains a controversial issue for clinicians.

Aims: To evaluate the basic characteristics of colorectal adenomas with malignancy risk using high-grade dysplasia (HGD) as an alternative marker for malignant transformation.

Methods: Data from Shanghai General Hospital between January 2017 and December 2021 were retrospectively analyzed. The primary outcome was the incidence of HGD in adenomas, which was used as a surrogate marker for the risk of malignancy. Odds ratios (ORs) for the HGD rate in adenomas were analyzed in relation to adenoma-related factors.

Results: A total of 9,646 patients identified with polyps during 57,445 screening colonoscopies were included in the study. Patients with flat polyps, sessile polyps, and pedunculated polyps represented 27.3% (N = 2,638), 42.7% (N = 4,114), and 30.0% (N = 2,894) of the total number, respectively. HGD was found in 2.41% (N = 97), 0.92% (N = 24), and 3.51% (N = 98) of sessile adenomas, flat adenomas, and pedunculated adenomas, respectively (P < 0.001). Multivariable logistic regression showed that polyp size (P < 0.001) but not shape (P > 0.8), was an independent predictor of HGD. Contrast to the diameter ≤1 cm, the OR value for diameters 1-2, 2-3, and >3 cm was 13.9, 49.3, and 161.6, respectively. The HGD incidence also increased in multiple adenomas (>3 vs. >1, ORs 1.582) and distal adenomas (distal vs. proximal adenomas, OR 2.252). Adenoma morphology (pedunculated vs. flat) was statistically significant in univariate analysis but not when size was included in the multivariate analysis. Besides, the incidence of HGD was also significantly higher in older patients (>64 vs. <50 years old, OR = 2.129). Sex (P = 0.681) was not statistically significant. All these associations were statistically significant (P < 0.05).

Conclusion: The malignant potential of polyps is mostly affected by their size but not by their shape. In addition, distal location, multiple adenomas, and advanced age were also correlated with malignant transformation.

Keywords: colonoscopy; colorectal adenomas; colorectal cancer; risk factors; screening.

FIGURE 1

Flowchart of the patients included in the study.

FIGURE 2

Histology: (A) Tubular, (B) tubulovillous, (C) villous, and (D) serrated adenomas.