br Fig Survival disparities of
Fig. 1. Survival disparities of CRC patients related to tumor location and epidemiological characteristics, (A)Drinker, (B) Nondrinker, (C) Smoker, (D) Nonsmoker, (E) Rural patients (F) Urban patients.
Fig. 2. Survival disparities of CRC patients related to tumor location and Clinico-pathological characteristics, (A) patients with T-stage I-II, (B)patients with T-stage III-IV, (C) patients without metastasis, (D)patients with metastasis, (E) patients with TNM stage I-II, (F) patients with TNM stage III-IV.
Comparison of risk mortality among colorectal cancer subsite location by epidemiological and clinico-pathological characteristics.
LCC vs. RCC
Rec vs. RCC
LCC vs. ReC
Abbreviation: CRC-colorectal cancer, RCC-right colon cancer, LCC-left colon cancer, ReC-rectum cancer, TNM-tumor-node-metastasis, FOLFOX-Folinic Doxorubicin Fluorouracil Oxaliplatin, FOLFIRI-Folinic acid Fluorouracil Irinotecan, OS-Overall survival, HR-Hazard Ratio, 95%CI-95%confidence interval. S. Abasse Kassim, et al.
Women patients from all combined subsite location showed a sig- nificant decrease in the risk for mortality compared to men
spectively) compared with nonsmoker or nondrinker. However, no significant diﬀerences in risk of mortality were observed for smoker vs. nonsmoker patients when stratified by LCC or ReC (All P > 0.05). In addition, patients with ReC revealed no significant risk of mortality between drinker vs. nondrinker (P = 0.348). (Tables 2 & S1)
for LCC vs. RCC in various subgroups including gender (female,
In this prospective study, we explored relevant factors aﬀecting the post-surgical survival prognoses of patients with CRC. Patients with LCC or ReC were most likely to have well/moderately diﬀerentiated grade, low depth of invasion, no regional lymph nodes, no distant metastasis and an earlier stage compared to those with RCC. Conversely, patients with RCC included a greater proportion of older patients and women, poorly diﬀerentiated carcinomas, higher T stage, higher distant metastasis and higher TNM stage compared to those with LCC. These findings are consistent with previous studies [15,16,18,21,32,33]. We also found that patients from rural areas were most likely to be smoker or drinker, consistent with findings in previous studies [34–36] in China. Interestingly, these results might serve as baseline to explain the worse prognoses in patients with RCC, as well as the survival disparity between urban and rural patients.
P = 0.012). Compared with this study, we found a slightly lower risk of mortality (1.46 for our study vs. 1.79 for Fangqi et al. ).This may be possibly due to factors such age (young subjects for our study vs. older subjects for Fangqi et al. ); lifestyle (smoking and drinking) which was not considered by Fangqi et al. ), as well as the sample size and distribution of subsite pathology among patients, which might be pointed as one of limitation for our study.
Our multivariate analysis revealed that age, N-stage and M-stage were found to be independent predictors of mortality for RCC, whereas tobacco smoking was found to be a predictor of mortality for LCC. Age, alcohol consumption, tumor grade, T-stage and treatment regimen were found to be predictors of mortality for ReC. These results are consistent with the findings of previous studies [22,33,38–42]. Jiang et al.  for example, used the surveillance, epidemiology and end results (SEER) database to determine the eﬀect of age on survival outcome in operated and non-operated patients. Of the 123, 356 patients with colon cancer, Noautonomous controlling elements found age to be an independent prognostic factor in stage I-IV of the disease and estimated that, at TNM stage I, older patients (41–80 years) were twice at risk of death (HR = 2.319, 95%CI (1.394–3.858), P = 0.001) compared with young patients (≤40 years). They also re-ported that, at the TNM stage IV, older patients had a 1.288-fold in-creased risk of mortality than the youngest (HR = 1.288, 95%CI (1.200–1.382), P < 0.001). Broadly, these findings were similar to our results (HR = 1.529, 95%CI (1.072–2.183), P = 0.019) for RCC pa-tients with age (> 60 vs. ≤60 years) at TNM stage I-IV. Also, Amri et al.  used 922 patients with colon cancer to evaluate the eﬀect of high grade disease on outcomes of surgery, and found that the colon cancer-related mortality doubled for patients with higher rate of nodal and distant metastasis, which is in accordance with our findings ((N-stage (N1 vs. N0), HR = 4.056, P = 0.012) and M-stage (M1 vs. M0, HR = 3.442, P < 0.0001)) for RCC. Regarding unhealthy behavior, our findings were in consonance with previous studies [5,43–45] and reinforce the importance of eliminating or reducing exposure to po-tentially modifiable risk factors such as cigarette smoking and alcohol consumption. Islami et al.  for instance, reported that more than half of all cancer deaths in men and one-third of cancer deaths in women in China in 2013 were attributable to the potentially modifiable risk fac-tors. Meanwhile, our study found that there were statistically sig-nificant diﬀerences in pairwise interactions between gender and either alcohol consumption, M-stage or TNM-stage for RCC. For LCC patients, significant diﬀerences in pairwise interactions were found between age and M-stage, gender and unhealthy behavior (smoke or drink), smoke and either residence or M-stage. Similar observations were found be-tween age and either smoking habits, tumor grade or M-stage, and for alcohol consumption with either smoking habits or M-stage for ReC patients. These findings are in line with previous studies [12,13,35,46,47]. For instance, Zeng et al.  used a population-based data from 17 cancer registries in China and found a great 5-year survival gap of CRC between urban patients (59.3%) and rural patients (52.6%) in China.