Variations in Demography and Prognosis by Colon Cancer Location

Discussion

The division of the colon into three different units for the study was derived out of surgical practice. The three most common procedures for colonic cancer are left and right colectomies and sigmoidal resection, which is often performed as a high anterior resection (1). Shorter and segmental resections are at times performed for various reasons, such as palliative surgery, or due to high patient comorbidity. Other conditions, such as multiple synchronous tumours, can require extended resections. Thus, the treatment and the surgical practice can support such a division into three entities rather than just that of a left and right spectrum. The results in this study do hint at differences between the left and the sigmoidal colon. A sigmoidal location was associated with male gender and better differentiation grades, which is in line with findings reported by Wray et al. (8). The differentiation grade itself is associated with other factors of the TNM system, including the risk of having node metastasis, which can affect staging and prognosis (9). There was also less emergency surgery than with location in the left colon itself. This may be related overall to emergency colorectal surgery with a high frequency of sigmoidal location but then mainly due to benign causes such as diverticulitis (10).

There are several differences in demography and pathology related to the location of the colonic tumour, such as age and gender distribution. There are also differences in survival. A right-sided location was the most frequent in this study, which concurs with other findings (11). A location in the right colon has been described in several large studies to carry a significantly worse prognosis (7, 12). This does partly concur with the results in this study but then only for OS. The lack of difference in CSS is interesting and there are probably several reasons for it. One is associated with the age of patients related to the location of tumour. A higher age often means a higher risk of death due to causes other than cancer, which will then translate into a difference between the two outcome parameters of OS and CSS (13, 14). Higher age is also associated with lower rates of inclusion in adjuvant chemotherapy programs for stage III disease. The sigmoidal tumours did confer the best overall survival prognosis (Figure 1). This is plausibly related to and concurs with the findings of better overall stages and better differentiation grades associated with the sigmoidal location. The lack of survival difference in stage I and II also indicate that the overall survival findings are related to pathology and staging rather other entities. In contrast, the left colon had a tendency to confer worse prognosis, possibly related to being a common location of obstruction and thus emergency surgery and high frequencies of metastatic disease as described by Nakagoe et al. and therefore more advanced stages of disease (15). The joint left colon disease data, with best and worst data, could then add up to better prognosis than that for the more frequent right-sided tumours, which could form a plausible explanation for the overall findings.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Overall survival by tumour location in colon cancer.

Biomolecular differences within the colon have been described, where MSI frequency is one example (16). This has also, at least in part, lead to the development of theories of colon cancer a heterogenic disease evolving along several pathways and with differing properties. There could be some implications from these findings. One aspect is for conducting and evaluating translational research. The possibility that some markers could be expressed differently in various parts of the bowel is interesting and may warrant further studies (17). A difference due to location could then be at risk of being misinterpreted as a difference due to tumour pathology. Another aspect is the evaluation of treatment including studies on new chemotherapies. The findings of Bleeker et al. of higher frequency of mutations in KRAS in tumours of the right colon could serve as an example (18). This would then concern both the outcome parameter and also the number of patients with each specific tumour location in each treatment arm. The latter could include both differences in pathology and possible differences in risk of chemotherapy related toxicity due to different location-associated tissue properties (19).

This study was retrospective and the number of patients is low in comparison to some of the referenced publications. However, the data is from a single centre where all the patients have been assessed and treated along the same principles. The surveillance and follow-up was carried out and recorded by the same staff. The population served by the hospital is also geographically defined, which facilitates the follow-up. We did, in many instances, have information on the cause of death and whether it was cancer-related or not, which provides the chance of using cancer-related survival as an outcome parameter. In cases of doubt, the individual case was recorded as being cancer-related for registration purposes. The pathology assessments were of good standard in meeting the UICC requirements and thus reducing the risk of stage migration phenomena, which also can add strength to the results (9). The findings support the hypothesis that the colon should be regarded as comprising at least three parts, where sigmoidal tumours have better pathology factors and thus survival prognosis. The findings could also challenge the hypothesis of right-sided tumours conferring a worse survival prognosis.