MATERIAL AND METHODS

Study design

A retrospective observational cohort study was conducted.

Study population

We studied 194 patients diagnosed with colorectal cancer (stages I-IV). All patients were over 18 years old and had been treated by the Oncology Unit of the Virgen de las Nieves University Hospital (HUVN) between 2000 and 2016. Patients with other types of colorectal cancer, synchronous neoplasms and those whose Clinical histories were incomplete were excluded. The study was approved by the HUVN Ethics and Research Committee, and was conducted according to the Declaration of Helsinki.

Clinical and sociodemographic variables

Clinical and sociodemographic data were collected from clinical histories: sex, diagnosis, date and age of diagnosis, surgery, tumor size, stage, ECOG (Eastern Cooperative Oncology Group) Performance Status, histopathological type, date of death and scheme of chemotherapy treatment. Tumor stage was determined according to the TNM classification of the AJCC (American Joint Committee of Cancer)(15).

Genetic variables

DNA was extracted from paraffin tumor samples from the Biobank of Andalusia (Collection Code: 19150007). The QIAamp DNA Mini Kit was used following the manufacturer’s instructions and stored at -40 ° C.
Gene mutations were analyzed by Real-Time PCR using TaqMan® probes:KRAS codons 12, 13 and 61 (rs121913529, rs121913530, rs121913535, rs112445441, rs17851045, rs121913238, rs121913240), NRAS codons 12, 13 and 61 (rs121913237, rs121434596, rs11554290, rs121913254),BRAF (rs113488022), PIK3CA exons 9 and 20 (rs121913273, rs104886003, rs121913281, rs121913279). Positive samples were confirmed by Sanger DNA Sequencing.

Outcome variables

Overall survival was the primary endpoint, and was defined as time from the diagnosis of CRC until death from any cause. The end date of the study was used when the death event had not occurred.

Statistical analysis

The normal distribution of the variables was determined using the Shapiro-Wilks test. Quantitative non-normally distributed variables were described by median and percentiles (25 and 75). Qualitative variables were expressed as frequencies.
The Kaplan-Meier method and the log-rank test were used to evaluate the influence of clinical, genetic and sociodemographic variables on survival (bivariate analysis). Multivariate Cox proportional hazard regression model was used to obtain the adjusted hazards ratio (HR) and 95% confidence interval (CI95%) for potential prognostic factors for survival.
The significance level for all the tests was p <0.05. The software used for the survival analysis was EZR 1.35 (R 2.3.0)(16).