Comparative evaluation of oncologic outcomes in colon cancer

PURPOSE: In this paper we report clinical variables on colon cancer series. Oncological outcomes were compared to low-income and high-income countries. METHODS: We analysed a prospective database of 51 colon cancer patients submitted to primary tumor resection between 2010 and 2011, showing clinical variables and oncologic outcomes. RESULTS: R0 resection obtained in 80.4%, 21.6% of patients was TNM stage IV, and only 13.7% showed TNM stage I. Diseasefree survival was 32 months, overall survival was 46 months, and the tumoral recurrence rate was 9.8%. Univariate analysis showed association of serum CEA levels ≥ 5 ng/dl (p= 0.004), presence of metastasis at diagnosis (p= 0.012), compromised surgical margins (p < 0.001) and poorer tumor differentiation (p= 0.041) to death. Multivariate analysis identified compromised surgical margins as an independent risk factor for death due to colon cancer (P=0.003; odds ratio=0.36; 95% confidence interval=0.004-0.33). Nowadays, 62.7% of patients are alive. CONCLUSION: Recurrence rate, disease-free survival and overall survival was similar to those observed in more developed countries. Serum CEA levels ≥ 5 ng/dl, the presence of metastasis at diagnosis, compromised surgical margins and poorer tumor differentiation were associated with death. A compromised surgical margin was the only independent risk factor for death.


Introduction
Colorectal cancer (CRC) is a significant health problem, and 1.4 million new cases were estimated worldwide, in 2012 1 .In the last decade, a progressive reduction in the annual incidence rate of the disease has been observed in the USA.However, CRC remains responsible for 50,000 annual deaths 2 .An increased incidence has been observed in patients under 50 years of age and in low-income countries 3 .In Brazil, almost 32,600 new cases were expected in 2014, and the disease represents the third leading cancer-related cause of death 4 .
CRC treatment has improved in the last decades, and overall survival has increased mainly due to the favorable effect of more efficient adjuvant chemotherapy regimens 5 .Despite the constant advances, the prognosis of CRC remains poor.Only 15% of patients with distant metastasis at diagnosis are expected to be alive after five years 6 .Five-year overall survival is less than 10% in Sub-Saharan Africa, 28% in India and 49% in Asia.On the other hand, high-income countries may reach a 65% overall survival after five years 7 .
Incidence and prognosis of distal and proximal colorectal cancers may not be the same.Differences could be explained by unequal molecular profiles characterized by allelic losses on chromosomes 17p, 18, and 5q, KRAS and p53 mutations observed more frequently in the left-side colon cancers [6][7] .Epidermal growth factor receptor (EGFR) is also overexpressed in distal colorectal cancers.Right-sided colon cancers are associated with v-RAF murine sarcoma viral oncogene homolog B (BRAF) mutations 8 .
The purpose of the present study is to describe the clinical outcomes in a group of CRC patients operated in a teaching hospital and to compare our results with recent data from other studies.

Methods
We conducted an analysis of a prospective database of colon cancer patients after approval from the Institutional Review Board.

Results
The study includes 51 patients.Mean age was 66 years (range, 34-88 years).Tables 1 and 2 summarize the clinical characteristics, histological features, and oncological outcomes of subjects.
The median follow-up was 46 months.Disease-free survival was 32 months (range, 0-60 months).Overall survival was 46 months (range, 1-62 months).Total mortality rate was 37.3% (n=19).Final mortality rate according to disease staging was 41.2% (n=7) in stage II, 25% in stage III and 72.7% in stage IV subjects.Figure 1  After, the final multivariate analysis was built and compromised surgical margins was identified as an independent risk factor for death due to colon cancer (P=0.003;odds ratio=0.36;95% confidence interval=0.004-0.33).Table 3 summarizes the univariate analysis of variable associated with recurrence and death.

Discussion
The incidence of CRC has decreased in developed countries.The same trend was not observed in low-income regions, which have been responsible for 50% of all new diagnoses of the disease, in the last 25 years.This observation may be associated with a "Western lifestyle" and risk factors such as smoking and excessive alcohol intake 9 .
The mean age at diagnosis was similar to that found in developed countries 10 .More than half of all subjects were diagnosed with a locoregional and metastatic disease.A possible explanation for this observation could be limited public health budget and low accessibility to screening programs 3,11 .
Although not evidenced in the present study, a lymphovascular invasion is usually associated with higher risk of lymph node spread, distant metastasis and poorer prognosis in other series 12 .The rate of poorly differentiated tumors may reach 25% in recent studies, and histological differentiation is considered an independent risk factor of poor prognosis 13 .We observed a higher risk of mortality in the poorly differentiated group; however, our results suggest that this may be due to the effects of confounding factors.In many series, the mucinous histologic subtype is associated with higher risk of incomplete resection and poor prognosis 14 .The high incidence observed in this series could be explained by geographical variations and the absence of rectal tumors that exhibit less mucinous differentiation 15 .We found no association of mucinous differentiation with recurrence and mortality, although the small number of patients may have interfered with our results.
Lymph node involvement is one of the most significant prognostic factors in CRC.Intergroup Trial INT-0089 evaluated 3,411 CRC patients and noted that the number of retrieved lymph nodes was related to overall survival, even in N0 patients 16 .A systematic review conducted by Chang et al. with 61,371 patients observed a positive association between overall survival 17 .The ideal number of retrieved lymph node should be at least 12.
However, recent studies have discussed the role of total numbers of lymph nodes retrieved, lymph node ratio, size and immune activation of lymph nodes 18,19 .
In the present study, the high number of R1 and R2 resections is probably a consequence of the high rate of advanced disease and palliative procedures.Of note, all stage IV patients who were submitted to primary tumor excision without metastasis resection were considered as R2.Compromised margins were an independent risk factor for mortality.
The surgeon should weigh several parameters before recommending a palliative procedure including patient preference, performance status, and symptoms.Moreover, some features of the tumor are important in this situation, such as the extension of the primary mass, tumor burden and response to systemic chemotherapy 20,21 .
Despite all treatment advances, tumor recurrence remains a major problem in CRC.In other series, recurrence rates may range from 4-16% 22,23 .Tumor perforation, advanced-stage disease, and poor differentiation appear to increase recurrence after curative treatment 24 .In the present study, no variable was associated with recurrence.However, this could be explained by the small sample size.

* Fisher's Exact Test
In the present study, overall survival was similar to the rates observed in more developed countries, as noted in Table 4.
Some factors were associated with death including serum CEA levels ≥ 5 ng/dl, a presence of metastasis at diagnosis, compromised surgical margins, and poorer tumor differentiation.However, a compromised surgical margin was the only independent risk factor for death.In summary, this results encourage us to achieve an R0 resection as a major goal of colon cancer therapy.

Conclusions
In the present study, recurrence rates and overall survival was similar to those observed in more developed countries.This could be explained by the fact that treatment and follow-up of patients were carried out in the wealthiest state in Brazil, with easier access to health resources.
Serum CEA levels ≥ 5 ng/dl, a presence of metastasis at diagnosis, compromised surgical margins and poorer tumor differentiation were associated with death.A compromised surgical margin was the only independent risk factor for death.
evidences a Kaplan-Meier survival curve according to pathological staging.A univariate analysis was performed, and no association was found between recurrence and age, serum CEA, lymph node involvement, a presence of metastasis at diagnosis, surgical margins, lymphovascular invasion, the presence of mucinous subtype and tumor differentiation.Death was associated with serum CEA levels ≥ 5 ng/dl, a presence of metastasis at diagnosis, compromised surgical margins and poorer tumor differentiation.
All subjects underwent surgery at Clinics Hospital, Survival curves were plotted using the Kaplan-Meier method.For all analyzes, a significance level of 5% was established.

TABLE 1 -
Main characteristics of patients with colon cancer.

TABLE 2 -
Main histological features and oncological outcomes of subjects with colon cancer.

TABLE 3 -
Univariate analysis -independent variables associated with disease recurrence or death

TABLE 4 -
Comparative evaluation of oncologic outcomes from low-income and high-income countries.