CHICAGO – A retrospective chart review of 1,218 cases at one institution suggests that nearly two-thirds of the colorectal cancers that occur in the interval between colonoscopies are lesions that were missed or incompletely removed during the previous colonoscopy.
"Operator-dependant variability with regards to first detection of colorectal cancer lesions, and effectiveness of treatment, may be a critical factors," according to lead author Chantal le Clercq, a medical student at Maastricht (the Netherlands) University Medical Center, who presented the data at the annual Digestive Disease Week.
All clinical and pathological records of all 1,218 patients diagnosed with colorectal cancer at Maastricht University Medical Center from Jan. 1, 2001 to Dec. 31, 2010 were reviewed. The patient population was 55% men and the mean age was 70 years. Patients with hereditary colorectal cancer, irritable bowel disease, or a previous history of colorectal cancer were excluded from the data analysis.
Cases also were excluded from analysis if records noted incomplete visualization of the colon, poor bowel preparation, or inadequate surveillance according to Dutch guidelines.
Interval colorectal cancers were defined as cancers occurring in the colon and rectum within 5 years of an index colonoscopy. All colorectal cancers were classified as either flat or protruded, and proximal or distal to the splenic flexure.
Interval colorectal cancers were defined as those developing from lesions missed because of a subtle microscopic appearance, such as flat tumors or those smaller than 1 cm, or from those incompletely removed in a previous polypectomy. When neither definition fit, the cause was ruled unclear.
Around 3% (39) of all colorectal cancers developed in patients who had a colonoscopy within the previous 5 years, said Ms. le Clercq. About one-third (11) were excluded from the analysis due to incomplete visualization or inadequate surveillance. About 2% (28) were interval colorectal cancers, which occurred mainly in men (71%) around 70 years old. On average, these cancers were diagnosed 26 months after the index colonoscopy.
The interval colorectal cancers were significantly smaller than noninterval cancers. Nearly 12% were smaller than 1 cm, 19% were 1-2 cm, and the remaining 69% were larger than 2 cm.
Interval colorectal cancers were more likely to be flat lesions than were non–interval colorectal cancers. The Tumor Node Metastasis (TNM) stage and histology were virtually identical. In both cases, more than 70% of procedures were performed by endoscopists.
Multiple logistic regression analysis, adjusted for age and sex, revealed that interval colorectal cancers were more often small (2.7 cm vs. 3.9 cm; odds ratio, 0.82; 95% confidence interval, 0.69-0.97; P = .020). They were more often flat than protruded (48% vs. 21%; OR, 3.75; 95% CI, 1.73-8.13; P = .001). They were more often located in the proximal than the distal colon (70% vs. 33%; OR, 5.07; 95% CI, 2.18-11.8; P less than .001).
"We found that 54% of cases (n = 15) could be explained by more subtle microscopic appearance. So, these tumors were less than 1 cm in size, or they had a flat appearance, and therefore, these could be explained as potentially missed lesions," said Ms. le Clercq.
Of the remaining 13 cases, 36% (n = 10) had no clear explanation, and nearly 11% (n = 3) could be attributed to incomplete polypectomy, because "they developed in the same [anatomical segment] as a previous polypectomy and therefore could be explained by potentially incomplete removed lesion," said Ms. le Clercq.
Ms. le Clercq said that 2.2% (1 of 45) of all colorectal cancers diagnosed at her hospital over the past 10 years were interval colorectal cancers. Around 64% of these lesions were related either to lesions that were overlooked, small, or flat, or were incompletely removed.
"These data highlight the need to train practicing endoscopists in the accurate detection and effective treatment of colorectal lesions, to prevent interval colorectal cancers in everyday practice," she said.
Ms. le Clercq disclosed no financial relationship with a commercial interest.