Superior survival in right-sided versus left-sided colon signet ring cell carcinoma

This study aimed to explore the association of tumor sidedness with the prognosis of patients with colon signet ring cell carcinoma (SRCC). Eligible patients were retrieved from the Surveillance, Epidemiology, and End Results database between 2004 and 2015. Cancer-specific survival (CSS) and overall survival (OS) were compared between patients with left-sided colon SRCC and those with right-sided lesions. A total of 2660 patients were included, among them, 1983 (74.5%) had right-sided colon SRCC. Compared to patients with left-sided colon SRCC, those who had the right-sided colon SRCC showed higher proportion of white race, female, aged ≥ 65 years, receiving total colectomy and ≥ 4 regional lymph node dissection; while had lower proportion of advanced AJCC stage. Besides, right-sided patients exhibited superior 5-year CSS (32.74% vs. 25.89%, P = 0.001) and OS (27.38% vs. 23.02%, P = 0.024) rates compared with left-sided ones. Multivariate analysis revealed that tumor sidedness was an independent prognostic factor. To be specific, patients with right-sided colon SRCC showed better CSS (HR: 0.873; 95% CI 0.777–0.981; P = 0.023) and OS (HR: 0.838; 95% CI 0.753–0.965; P = 0.002). Moreover, subgroup analysis demonstrated superior CSS and OS for right-sided patients in most subgroups. Tumor sidedness was an independent prognostic indicator for colon SRCC. Besides, patients with right-sided colon SRCC have superior prognosis than those with left-sided lesions.

The endpoint of this study included cancer-specific survival (CSS) and overall survival (OS). CSS was defined as the period from diagnosis to death attributed to colon SRCC. OS was defined as the period from diagnosis to death of any cause.
Statistical analyses. Kaplan-Meier (K-M) method was employed to estimate the survival, followed by log-rank test for assessing the differences of CSS and OS. Notably, if variables had P values ≤ 0.1 in univariate analysis, they were incorporated into multivariate Cox proportional hazard analysis. Before multivariate analysis, Schoenfeld residual method was used to test the proportional risk assumption of Cox regression model, which revealed that the proportional risk assumption of this study was tenable. Similarly, Cox regression analysis was also used for stratified analysis. SPSS software (SPSS Inc., Chicago, USA, version 19.0) was used for statistical analysis, and GraphPad Prism 5 was utilized for plotting survival curves and generating forest plots. A twosided P < 0.05 indicated statistical significance. www.nature.com/scientificreports/ and 2015, with a median follow-up of 16 months (range 0-155 months). Then, these subjects were classified into left-sided (n = 677, 25.5%) and right-sided (n = 1983, 74.5%) colon group. The detailed procedure of patient selection was presented in Fig. 1. Table 1 showed the baseline characteristics of included patients divided based on tumor sidedness. According to the results, differences in year at diagnosis (P = 0.049), insured status (P = 0.001), age (P < 0.001), gender (P < 0.001), race (P < 0.001), AJCC stage (P < 0.001), surgery (P < 0.001), LNDs (P < 0.001), chemotherapy (P < 0.001) and radiotherapy (P < 0.001) between two groups were statistically significant. Prognostic effect of tumor sidedness on patient. Differences in CSS (P = 0.001) and OS (P = 0.024) between right-sided and left-sided groups were statistically significant (Fig. 2) Subgroup analysis of the effect of cancer sidedness on CSS and OS. The impacts of tumor sidedness on patient CSS and OS were determined based on subgroup analysis. As indicated by forest plots for subgroup analysis, right-sided patients had superior OS and CSS than left-sided patients in most subgroups (Figs. 3 and 4).

Discussion
To our knowledge, this is the first study to explore the effect of tumor sidedness on the survival of colon SRCC patients. In order to make our data more convincing, patients were strictly screened. First of all, only pathologically confirmed primary colon SRCC patients were selected. Then, patients with secondary primary tumor were deleted. Finally, we also deleted patients with missing important clinicopathological variables, such as AJCC stage, location, etc. Finally, a total of 2660 colon SRCC patients were enrolled into analysis. And our results indicated that right-sided patients were associated with better survival and prognosis. After adjusting for tumor and demographic features, the cancer-specific death and overall death in patients with right-sided colon SRCC were reduced by 12.7% and 16.2%, respectively, relative to those in left-sided patients. In general, tumor sidedness serves as an independent factor to predict prognosis among colon SRCC patients. As stated in studies, right-sided CRC patients are associated with older age, female, greater tumor size, more advanced tumor stages, and poor tumor differentiation 10,26-28 . These are partly consistent with our findings, except that we found that right-sided colon SRCC tended to have earlier stage. www.nature.com/scientificreports/ www.nature.com/scientificreports/ At present, no uniform results have been obtained for the prognosis of left-sided versus right-sided CRC, and more studies revealed poorer survival in right-sided primary tumor [29][30][31] . For instance, Petrelli et al. performed a systematic review and meta-analysis by recruiting 66 studies, and further suggested that left-sided CRC was correlated with the decreased mortality rate (HR: 0.82; 95% CI 0.79-0.84) 29 . In addition, by enrolling stage III CRC patients (n = 1,869), Taieb et al. discovered that right-sided cancer patients showed worse OS (HR: 1.25; 95% CI 1.02-1.54) compared left-sided tumor patients 30 . Another study revealed the association of left-sided tumor with better survival in patients with wild-type RAS metastatic CRC 31 .
However, different outcomes have also been suggested by other studies. By adopting propensity score matching, Rene Warschkow demonstrated that right-sided CRC patients had superior OS (HR = 0.92, 95% CI 0.8-0.94, P < 0.001) together with better CSS (HR = 0.90, 95% CI 0.87-0.93, P < 0.001) 32 . Generally speaking, no consensus has been reached about the prognosis for left-sided vs. right-sided CRC. In consideration of the numerous heterogeneities (such asinclusion or exclusion criteria, study design), it remains difficult to compare those available studies.
Notably, the SEER cancer registry is beneficial because it allows us to evaluate rare cancers in a great crosssection and offers the long-term follow-up data with no intrinsic institutional bias. Nonetheless, certain limitations should be noted in our study. Firstly, 2303 (46%) patients were excluded because of the strict screening criteria in this study, which may lead to certain selection biases, but it is inevitable in retrospective studies 17,19 . Secondly, not all required data are contained in the SEER database to comprehensively answer our questions. For instance, there is no available information regarding family history, performance status, extramural vascular invasion or obstruction/occlusion status. Thirdly, the variables of chemotherapy and radiotherapy showed low sensitivity. Only a part of patients definitely received radiotherapy or chemotherapy, which is defined as "yes"; while others are uncertain, which is defined as "no/unknown". Therefore, interpreting this part of data should be extremely careful. Finally, several important prognostic information is unavailable from the SEER database, such as Microsatellite stability/Microsatellite instability (MSS/MSI) status. Although it is better to obtain more details, we believed that the present available data from SEER database could fit our research objectives very well. The above concerns should be investigated in future studies.

Conclusions
In conclusion, our present findings suggest that tumor sidedness could serve as an independent factor to predict prognosis in patients with colon SRCC. Compared with patients with left-sided colon SRCC, the right-sided patients have superior prognosis. More in-depth research is warranted to further examine tumor heterogeneity www.nature.com/scientificreports/ together with the related biological factors. In addition, tumor sidedness may be identified as an independent factor when selecting the therapeutic regimen in colon SRCC management. Collaborations are valuable for merging data derived from multiple institutions in the future.