The Malignant Potential of Freshly Developed Colorectal Polyps According to Age Free

It has been generally accepted that colorectal cancers and adenomas are found more frequently in older population (1-5). On the basis of the adenoma-carcinoma sequence theory, colorectal adenomas are regarded as precancerous lesions that progress to colorectal cancers, taking a long time in multistep malignant transformation (6-10). Based on this hypothesis, the oldest polyp, which developed at the earliest age, has the greatest chance of malignant transformation. In older people, colorectal polyps may include those that developed a long time ago, having enough time for malignant transformation, leading to higher incidence of cancers among them. Thus, the reason why colorectal cancers are more prevalent in older people can be explained.

However, the polyps developing at old ages may have a higher intrinsic potential for malignant transformation, requiring less time for cancer development from benign adenomas; or there may be another way of carcinogenesis in older people other than adenoma-carcinoma sequence (4, 11-14).

To examine these possibilities, we conducted the present study. Because the polyps found at one-time colonoscopy are a mixture of lesions that developed at various earlier times, we analyzed newly developed adenomas after clearing all neoplastic lesions by total colonoscopies.

We consecutively enrolled asymptomatic Japanese who underwent total colonoscopy as part of their annual medical health checkup at the Kameda General Hospital or Makuhari Clinic between 1988 and 2002. None had overt rectal bleeding, recent change in bowel habits, or recent or current lower abdominal pain. They had not been screened by fecal occult blood tests or other examinations for colorectal cancer, and they underwent colonoscopy as the first-choice examination to check for colorectal diseases. Subjects with a personal history of colorectal cancer, colorectal polyps, or inflammatory bowel disease had been excluded. All subjects were encouraged to undergo annual follow-up examinations by total colonoscopy irrespective of the results of previous examinations. Subjects with incomplete intubation or incomplete removal of polyps were excluded for analysis at the initial examinations. Subjects with invasive cancers found at the initial examinations underwent surgical operation and were excluded from follow-up analysis. All colorectal neoplasms found at follow-up colonoscopies were included in the analysis. The subjects enrolled in this study included the subjects of a previous study (5).

The protocol was approved by the ethics committee of each institution, and informed consent was obtained from each subject according to the Declaration of Helsinki.

Each subject ingested 2 liters of polyethylene glycol electrolyte solution for bowel preparation. Medium-length colonoscopes were used (mainly PCF200 and PCF230, Olympus, Tokyo, Japan). During the procedure, the location and size of each polyp were documented. The size of each polyp was estimated with open-biopsy forceps (diameter 8 mm). All of the polyps ≥5 mm and lesions <5 mm suspected of neoplasia were resected endoscopically or by surgical operation. We did not remove small polyps that were felt to be hyperplastic or inflammatory by macroscopic endoscopic examination. All pathologic specimens were evaluated by the same pathologists of the Kameda General Hospital. Neoplastic lesions were classified into invasive cancer, adenoma with high-grade dysplasia, and low-grade dysplasia.

Malignant lesions were defined as invasive cancer or high-grade dysplasia. The proportion of malignant lesions among all neoplastic lesions was calculated for lesions found at the follow-up colonoscopies stratified by the age of patients, the size and location of colorectal lesions, or the previous findings at the initial colonoscopy.

The subjects were divided into the young (<50 years old), middle (50-59 years old), and old (≥60 years old) age groups. The size of the lesion was classified into <10, 10 to 19, and ≥20 mm. The location was classified into right-side or left-side colon divided at the splenic flexure, which was included in the left-side colon. The findings at the initial colonoscopy were classified according to the most advanced lesion as “no neoplasm,” with no neoplastic lesions; “small adenoma,” with only benign adenoma of <10 mm; and “advanced lesion,” with adenoma of ≥10 mm or malignant lesion.

Analysis regarding the proportion of malignant lesions with respect to age was done using the χ² trend test or exact trend test. When the proportion was compared between two groups, χ² test or Fisher's exact test was used. A multivariate analysis of associating factors for malignant lesions was conducted by a multiple logistic regression model. A two-sided P value of <0.05 was considered statistically significant. Analyses were done with SAS for Windows software version 8.2 (SAS Institute, Inc., Cary, NC).

A total of 68,053 screening colonoscopies were done from January 1988 to March 2002, including first-time endoscopies for 21,419 subjects, where all neoplastic lesions, if any, were endoscopically removed. Out of them, 11,912 subjects underwent at least one follow-up colonoscopy: 3,276 once, 2,115 twice, and 6,521 thrice or more, amounting to a total of 44,065 follow-up colonoscopies, with the mean follow-up period of 5.47 years, the mean examination number of 4.71 times, and the mean examination interval of 1.47 years.

At the follow-up colonoscopies on 11,912 subjects, a total of 8,271 colorectal neoplasms were detected in 2,736 (23.0%) subjects. Out of them, 41 lesions (0.50%) were malignant, including 9 invasive cancers (Table 1). The proportion of malignant neoplasms among all the neoplastic lesions was 0.35% (10 of 2,885), 0.31% (11 of 3,509), and 1.07% (20 of 1,877) in the young, middle, and old age groups, respectively, revealing a significant trend of increase with age (P = 0.002).

The proportion of malignant neoplasms stratified by the size is shown in Table 2, showing a strong association between the two (P < 0.0001). When stratified also by age, the proportion increased with age in each size group (Table 2). The proportion of malignant neoplasms among lesions <20 mm was 0.21%, 0.31%, and 0.75% in the young, middle, and old age groups, respectively, showing a significant increasing trend with age (P = 0.005).

Out of 8,271 colorectal neoplasms, 4,304 lesions (52.0%) were located on the right-side colon, and 3,967 were located on the left-side colon. The proportion of malignant neoplasms in each location group is shown in Table 3. The proportion of malignant neoplasms tended to increase with age in each location group. The proportion of malignancy was usually greater on the left-side colon than on the right-side colon (P = 0.0001). However, it was similar on both sides in the old age group due to its steep age-related increase on the right-side colon.

Then, the 11,912 subjects were classified according to the most advanced lesion found at the initial colonoscopy, into no neoplasm group (n = 9,409, 79.0%), small adenoma group (2,103, 17.7%), and advanced lesion group (400, 3.4%). Out of 8,271 colorectal neoplasms detected during follow-up colonoscopies, 4,584 lesions (0.49 per person) developed from the no neoplasm group, 2,955 lesions (1.41 per person) developed from the small adenoma group, and 732 lesions (1.83 per person) developed from the advanced lesion group. A greater number of colorectal neoplasms developed from the subjects with more advanced lesions at the initial examinations (P < 0.0001). The proportion of malignancy was 0.54% (25 of 4,584), 0.41% (12 of 2,955), and 0.55% (4 of 732) in no neoplasm, small adenoma, and advanced lesion groups, respectively, with no overall difference found among the groups (Table 4). The proportion of malignancy was 1.07% (15 of 1,401), 0.51% (5 of 977), and 0.80% (2 of 249) in each group at the first surveillance, also with no overall difference found among the groups.

However, when stratified by age, the proportion of malignant neoplasms tended to increase according to age groups in each group of the initial lesions. The trend was shown in the no neoplasm group (0.43%, 0.36%, and 1.28%, in the young, middle, and old age groups, respectively, P = 0.04) and in the small adenoma group (0.11%, 0.33%, and 0.81%, P = 0.02).

A multivariate analysis controlling age, gender, examination interval, follow-up time, initial number of polyps, initial maximum size of polyps, and family history of colorectal cancer was conducted. Among these variables, age was the only significant factor for the malignant character of polyps (P < 0.001 by multiple logistic regression model; Table 5).

Colorectal cancers are presumed to develop from colorectal adenomas after a long-term process of malignant transformation (6-10). Colorectal polyps discovered at a one-time colonoscopy may be chronologically heterogeneous, having arisen at various earlier ages. In the present study, we analyzed newly developed adenomas after the confirmation of the absence or complete removal of colorectal neoplasms by total colonoscopy to evaluate the malignant potential of polyps in relation to the actual age at which they developed.

The present study showed higher malignant potential of colorectal neoplasms in the old than in the young. This trend was shown even when adjusting for the size of polyps. The size is well known to be associated with the malignant character of colorectal polyps (15-17). It may suggest that the polyps on the old people are more malignant than the polyps with similar appearance on the young people. Although the trend was obscure in the lesions ≥20 mm, they might be particularly atypical lesions or missed lesions.

The malignant tendency of polyps in old people was generally stronger on the left-side colon than on the right-side colon, which is compatible with the prior literature of cross-sectional studies (17). However, at the follow-up examinations, the malignant potential of neoplasms on the right-side colon increased substantially in the old to reach the same level as that on the left-side colon. It may suggest the accelerated adenoma-carcinoma sequence or another type of carcinogenesis in the old age on the right-side colon. This may be related with our finding and those of others that the incidences of right-side colon cancers and adenomas increase with age (4, 18-20).

High malignant potential on the right-side colon resembles the cases in hereditary nonpolyposis colorectal cancer. Rijcken et al. (21) reported that adenomas in hereditary nonpolyposis colorectal cancer patients had increased susceptibility to malignant conversion than sporadic adenomas found in control patients, specifically on the right-side colon. Mechanisms similar to hereditary nonpolyposis colorectal cancer may be engaged in carcinogenesis in the right-side colon of old people (22-25).

Another notable result of this study was that the malignant potential was not associated with the lesions found at the initial colonoscopy. We and others (5, 26, 27) have reported higher incidence of metachronous colorectal adenomas in patients with larger or more advanced colorectal lesions at the initial colonoscopy. A higher incidence of overall neoplasms from the subjects with more advanced lesions at the initial examination was also shown in this study. However, the malignant potential in each polyp was not associated with the initial lesions, but highly associated with age.

The limitation of the present study is that the malignant lesions found in this study might be extraordinary, rapid-growing ones because of short intervals between colonoscopies. The question whether the majority, more slowly growing polyps, also had higher malignant potential in older people should be addressed by long-term observation or other investigational method in the future.

In conclusion, we showed in this study that the development of malignant lesions in small sizes increased on the colon or rectum at older ages. The polyps found in the old should be treated more carefully than the polyps with the same size in the young.