Comparison of survival between right‐sided and left‐sided colon cancer in different situations

Abstract Mountain of studies has showed that right‐sided colon cancer (RSCC) and left‐sided colon cancer (LSCC) have different clinical presentation and biologic features and should be considered as two distinct disease entities. The survival difference between RSCC and LSCC remains controversial. Using Surveillance, Epidemiology, and End Results (SEER) database, we identified colon adenocarcinoma patients from 2004 to 2013. The 5‐year cause‐specific survival (CSS) was our primary endpoint. All statistical analyses were performed using the Intercooled Stata 13.0. All statistical tests were two‐sided. The study included 95,847 (58.72%) RSCC and 67,385 (41.28%) LSCC patients. RSCC patients were older, more often females, more Caucasian, more unmarried, more advanced T and N stage, larger tumor sizes, and more poorly differentiated tumor, while LSCC patients had more stage IV diseases. Location was an independent prognostic factor in the multivariable analysis. Compared with RSCC patients, the hazard ratio for LSCC was 0.87, 95% CI: 0.85–0.89 P < 0.001. There was no survival difference between RSCC and LSCC in the following situations: older than 68 years old, T3–4, N0, poorly differentiated, and undifferentiated diseases. We firstly reported that RSCC patients had a better prognosis than LSCC in mucinous adenocarcinoma/signet ring cell carcinoma patients. RSCC patients also had a better prognosis than LSCC in stage II disease. There is a need for further subdivisions when analyzing the survival difference between RSCC and LSCC patients. RSCC had lower mortality rate than LSCC in stage II disease and mucinous adenocarcinoma/signet ring cell carcinoma patients.


Introduction
Colon cancer is among the leading causes of cancer-related deaths all over the world [1,2]. Mountain of studies has showed that right-sided colon cancer (RSCC) and left-sided colon cancer (LSCC) should be considered as two distinct disease entities [3][4][5][6]. Differences in embryologic development, clinical presentation, patient demographics, and tumor biology between RSCC and LSCC have been clearly reported in the literatures [3,[5][6][7][8][9]. It is well known that RSCC arises from the embryonic midgut and is perfused by the superior mesenteric artery, while LSCC originates from the hindgut and is served by the inferior mesenteric artery [6]. Moreover, the capillary network surrounding the LSCC is multilayered, whereas that of the LSCC is single-layered, possibly relating to the greater water absorption and electrolyte transport capacity of the former [10]. The difference in anatomic structure may partly explain the different clinical presentation between RSCC and LSCC, such as more advanced T stage with severe symptoms (passage trouble or abdominal mass) in RSCC patients [7,11].
Whether the biologic and clinical differences between RSCC and LSCC have translated into clinically meaningful prognostic difference is still controversial. Although accumulating evidences suggest that RSCC patients have a worse prognosis than LSCC patients [4,5,[12][13][14][15][16], Weiss JM et al. [15] used the Medicare beneficiaries of colon adenocarcinoma to compare survival between RSCC and LSCC patients by stage and found that there was no overall difference in 5-year mortality between RSCC and LSCC patients. Their further analysis showed that stage II RSCC had lower mortality, while stage III RSCC had higher mortality than LSCC [15]. Except for stage, are there other factors affecting the survival comparison between RSCC and LSCC patients? In this study, we used data from Surveillance, Epidemiology, and End Results (SEER) cancer registry program of individuals diagnosed with colon adenocarcinoma from 2004 to 2013 to compare the survival and clinicopathologic features between RSCC and LSCC patients in different situations.

Statistics
The patients' demographic and tumor characteristics were summarized with descriptive statistics. Comparisons of categorical variables between right and left colon cancer patients were performed using the chi-squared test, and continuous variables were compared using Student's t test. The primary endpoint of this study was 5-year cause-specific survival (CSS), which was calculated from the date of diagnosis to the date of cancer-specific death. Deaths attributed to colon cancer were treated as events, and deaths from other causes were treated as censored observations. Survival function estimation and comparison between RSCC and LSCC were performed using Kaplan-Meier estimates and the log-rank test. The independence of the prognostic effect of location was evaluated by adjusting for other known factors including age at diagnosis and tumor stage. The multivariate Cox proportional hazard model was used to evaluate the hazard ratio (HR) and the 95% confidence interval (CI) for all the prognostic factors. All of statistical analyses were performed using the Intercooled Stata 13.0 (Stata Corporation, College Station, TX). Statistical significance was set at two-sided P < 0.05.

Database
The SEER database is the largest publicly available cancer dataset. It is a population-based cancer registry across several disparate geographic regions. The SEER research data include cancer incidence and prevalence as well as demographic information tabulated by age, sex, race/ethnicity, year of diagnosis, and geographic region. The dataset we used for this analysis was Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) Research Data .

Patient population
The study population was based on the SEER cancer registry. Since the American Joint Committee on Cancer (AJCC) 7th Tumor-Node-Metastasis (TNM) staging system was released in 2010 and if we used this staging system, there would be no 5-year survival due to insufficient followup, so we picked up the AJCC 6th TNM staging systems. Meanwhile, since the AJCC 6th TNM staging system was released in 2004, we selected patients from 2004 to 2013.
All patients had active follow-up and the survival month was over 1 month. Patients were excluded if they had more than one primary cancer, but colon cancer was not the first one or had unknown cause of death. AJCC 6th TNM staging systems were used for the staging. We excluded patients whose TNM stage was unknown.

Patient baseline characteristics
The study identified 163,232 colon adenocarcinoma patients including 80,599 (49.38%) men and 82.633 (50.62%) women. Of these patients, 95,847 (58.72%) were RSCC and 67,385 (41.28%) were LSCC. The mean age of the whole population was 67.28 ± 13.61 (Mean ± SD) with a median age of 68 years old. Table 1 showed the basic features between these two groups of patients. The proportion of men was significantly higher in LSCC than in RSCC patients, P < 0.001. The median age of LSCC patients was significantly younger than RSCC patients, 64 and 71 years old, respectively, P < 0.001. More LSCC patients were married. For the TNM stage, LSCC patients had higher percentage of stage I and IV diseases. Poorly differentiated and undifferentiated adenocarcinoma or mucinous adenocarcinoma/signet ring cell carcinoma was less common in LSCC than in RSCC patients. LSCC was more likely to be detected at a smaller tumor size than RSCC patients (median tumor size: 40 mm vs. 45 mm), P < 0.001. RSCC patients received more surgery and less radiation than LSCC patients. For those receiving operation, RSCC had more lymph nodes resected and fewer positive lymph nodes.
The median age of the whole population was 68 years old. We therefore divided the patients into two groups according to the age: <69 years old (younger patients) and >68 years old (older patients). The younger patients had a significantly better 5-year CSS than the older patients (71.2% vs. 67.2%, P < 0.001; Table 2).
No doubt, the TNM stage was significantly correlated with survival. The 5-year CSS was 93.6%, 84.8%, 68.3%, and 13.1% for patients from stage I to stage IV, respectively, P < 0.001. For the histology subtypes, signet ring cells had worse 5-year CSS than the mucinous adenocarcinoma and other adenocarcinoma. When we analyzed the 5-year CSS in patients with different grades, we found that the survival became poorer as the tumor grades progressed from well to undifferentiated, 83.0% for well differentiated, 72.8% for moderately differentiated, 55.9% for poorly differentiated, and 54.2% for undifferentiated tumors, P < 0.001.

Multivariate analysis
Variables showing a trend for association with survival (P < 0.05) were selected in the Cox proportional hazards model. Age, married status, ethnicity, location, TNM stage, histologic subtypes, grade, tumor size, as well as surgery were all independent prognostic factors in the multivariable analysis. Compared with RSCC patients, the HR for LSCC patients was 0.87, 95% CI: 0.85-0.89, P < 0.001 (Table 3).

Survival difference between RSCC and LSCC in different situations
We further compared the survival difference between RSCC and LSCC in different situations (Table 4). We found that LSCC patients had better prognosis than RSCC in both men and women, younger patients, all ethnicity subgroups, different married status, well and moderately differentiated adenocarcinoma patients and also in all the There was no significant difference between RSCC and LSCC in older patients or in those with poorly differentiated or undifferentiated adenocarcinoma. Meanwhile, no survival difference between RSCC and LSCC in T2, T3 or N0 disease was found.
For patients with different TNM stages, LSCC had a better prognosis than RSCC except for stage II disease (Fig. 1). Surprisingly, the survival trend reversed in stage II disease, 83.7% and 85.5% for LSCC and RSCC patients, respectively, P < 0.001. As we know that when the resected number of lymph nodes was <12, it may lead to  inappropriate staging, especially for stage II disease. To better understand the survival difference in stage II disease, we analyzed the survival difference between RSCC and LSCC when resected lymph nodes were less 12 and over 11, respectively (Fig. 2). We found that there was no significant survival difference between these two groups when the resected number of lymph nodes was less 12, P = 0.7829 ( Fig. 2A). When the resected number of lymph nodes was over 11, stage II RSCC patients had significantly better prognosis than LSCC patients, P = 0.0228 (Fig. 2B).
Moreover, LSCC patients also had a poorer survival than RSCC when the histology subtypes were mucinous adenocarcinoma or signet ring cell carcinoma (Fig. 3).

Clinicopathologic features of stage II patients between RSCC and LSCC patients
To understand why LSCC patients had poorer survival than RSCC in stage II diseases, we compared the clinicopathologic features of stage II patients between RSCC and LSCC (Appendix S1). Overall, the clinicopathologic features between RSCC and LSCC were similar in stage II disease and in the whole population. Except that LSCC patients had more T4 diseases than RSCC in stage II while in the whole population, LSCC had less T4 diseases.

Clinicopathologic features of mucinous adenocarcinoma/signet ring cell carcinoma patients between RSCC and LSCC patients
Similarly, we compared the clinicopathologic features of mucinous adenocarcinoma/signet ring cell carcinoma patients between RSCC and LSCC patients (Appendix S2). Compared with the whole population, more RSCC patients had stage I disease in mucinous adenocarcinoma/signet ring cell carcinoma patients.

Discussion
A number of studies have been carried out in different regions of the world to describe the differences between RSCC and LSCC. Regarding the difference in biologic behavior and clinical presentation, RSCC and LSCC were suggested to be considered as two disease entities [5,9,15,17]. In this study, we found that the relationship between clinicopathologic features and tumor location in colon cancer was not straightforward. Specifically, RSCC patients not only had some adverse features, such as older, more unmarried, more advanced T and N stage, larger tumor sizes, and more poorly differentiated tumor which were similar to the previous reports [6,11,13,15,18,19], but also had some good features, including less metastasis diseases and fewer numbers of positive lymph nodes. Most previous comparisons of clinicopathological features between RSCC and LSCC only included stage I-III diseases and they concluded that RSCC had more advanced stages [5,15,[20][21][22]. Here, we pointed out that actually RSCC patients had less metastasis diseases than LSCC. The complicated relationship between clinicopathologic features and tumor location in colon cancer might partly explain the controversial results of survival comparison between RSCC and LSCC patients [4,5,[12][13][14][15][16]. Some found that RSCC had better survival than LSCC [11,[18][19][20]23]. Other studies considered that location of colon had no relationship with survival [15,24]. Thinking about the controversial results in the literatures, we hypothesized that the comparison of survival between RSCC and LSCC might vary in different situations.
Here, we firstly reported that RSCC had better prognosis than LSCC in mucinous adenocarcinoma/signet ring cell carcinoma patients. Except for this, RSCC also had better prognosis than LSCC in stage II diseases when the number of resected lymph nodes was over 11, consistent with previous reports [15,19,21]. To further understand the above findings, we compared the clinicopathologic features between RSCC and LSCC in stage II patients and mucinous adenocarcinoma/signet ring cell carcinoma patients. We found that the clinicopathologic features were similar in the subgroups and in the whole population. It seemed that the better survival of RSCC patients in the above two situations was more likely related to tumor biology. Previous study showed that survival in stage II/III colorectal cancer was independently predicted by microsatellite instability (MSI), but not by specific driver mutations [25]. MSI is predominantly seen in RSCC (about 25%) [3], while <5% in LSCC [6]. Mucinous adenocarcinoma was more common in RSCC and was reported to have more MSI than nonmucinous adenocarcinoma [26]. High MSI is related to a better overall survival [27][28][29] despite the fact that the effect of adjuvant chemotherapy, especially   [30]. We hypothesized that MSI was the major contributor to the reverse mortality between RSCC and LSCC in stage II and mucinous adenocarcinoma/signet ring cell carcinoma patients. Further researches are needed to confirm our hypothesis.

5-fluorouracil, is reduced in patients with MSI
In the multivariate analysis, we found that location was an independent prognostic factor in the whole population. LSCC had lower mortality rate than RSCC with a hazard ratio of 0.87. There were other studies using the SEER database or SEER-Medicare database trying to explore the role of location on survival. In Weiss JM's study, they found no difference in 5-year mortality between RSCC and LSCC patients [15]. Their study was limited in the patients' age and stage. All the patients were 66 years and older and they only enrolled stage I to III patients. LSCC patients were younger and had more stage IV diseases [13]. In our studies, we showed that LSCC had better prognosis than RSCC in younger patients and also stage IV patients. After excluding patients whose prognosis favoring LSCC, it was not hard to understand why no survival difference was found in Weiss JM's study. Meguid et al. [19] analyzed patients who underwent surgical resection for invasive colon adenocarcinoma using the SEER database between 1988 and 2003, and found that RSCC had worse prognosis than LSCC patients, which was similar to our result. This study was also limited to patients' selection. Only patients who received surgical resection were considered. The patients' selection in our present study was more close to real world.
Potential limitations of our study should be taken into consideration. Unmeasured factors in SEER database, such as chemotherapy and tumor biology, including MSI status might play roles in patient outcome. Recent reports showed that RSCC and LSCC even had different response to the anti-Epidermal growth factor receptor (EGFR) and anti-Vascular endothelial growth factor (VEGF) monoclonal antibody [31]. We could not fully evaluate the impact of chemotherapy and target therapy on survival of RSCC and LSCC patients.
In conclusion, the relationship between survival and tumor location in colon cancer was not straightforward. There is a need for further subdivisions when analyzing the survival difference between RSCC and LSCC. We found that RSCC patients had better prognosis than LSCC in stage II disease or mucinous adenocarcinoma/signet ring cell carcinoma patients.