Proximal adenoma location did not predict high-grade dysplasia in a large registry study.
In fact, the odds of high-grade dysplasia were about 25% lower for proximal versus distal adenomas (odds ratio, 0.75), reported Thomas Rösch, MD, of University Hospital Hamburg-Eppendorf, Hamburg, Germany, and his associates. A third of adenomas in the study lacked location data, but in sensitivity analyses, the odds of high-grade dysplasia fell to 0.72 when these lesions were assumed to be proximal and rose to 0.96 when they were assumed to be distal.
Interval colorectal cancers probably are more likely to be proximal than distal because of a “combination of endoscopy-related factors and biology,” not because of histologic differences alone, the researchers wrote. The report was published in Clinical Gastroenterology and Hepatology.
Interval cancers are more common in the right colon, as several studies have noted. However, it was unclear whether this phenomenon represented a higher miss rate, a lower rate of successful polypectomy, or an increased risk of malignant histology in the proximal colon, the researchers wrote. Accordingly, they analyzed data on 594,614 index adenomas detected during more than 2.5 million screening colonoscopies performed between 2007 and 2012 and entered into the German National Screening Colonoscopy Registry.
A total of 3.5% of index adenomas showed high-grade dysplasia, which correlated most strongly with larger size, said the researchers. In fact, the odds of high-grade dysplasia were 10-fold higher when index adenomas measured at least 1 cm than when they were smaller. High-grade dysplasia also was significantly more frequent when patients were older than 64 years, were male, and when they had pedunculated versus flat lesions. Given the large size of the dataset, all these associations were statistically significant.
Sessile lesions were slightly more likely to be high-grade compared with flat lesions, the investigators noted. Many proximal interval cancers arise from sessile serrated polyps, which may be subtle and difficult to detect or to resect completely, they continued. At the same time, colonoscopy also might be more likely to miss flat, serrated lesions when they are located proximally, and these lesions can become more aggressive over time. Thus, “[e]ndoscopist factors, such as missed lesions or incompletely removed lesions, may account for the predominance of proximal interval colorectal cancers.”
Like other registry studies, this study lacked uniform histopathologic definitions or central histopathology review. The dataset also covered only the largest or most histologically remarkable adenoma for each patient. However, the study findings did not change substantially after the researchers controlled for patients with missing location data, which presumably included patients with multiple polyps in both proximal and distal locations.
The researchers did not disclose external funding sources. They reported having no conflicts of interest.
SOURCE: Rösch T et al. Clin Gastroenterol Hepatol. 2018 Jun 11. doi: 10.1016/j.cgh.2018.05.043.
