Investigating clinical and molecular characteristics of patients with PIK3CA mutation with and without KRAS mutations in colorectal cancer.

201 Background: PIK3CA mutations are some of the most common somatic mutations in colorectal cancer (CRC) and can be seen concurrently with KRAS mutations. However, it is unknown whether the presence of a concurrent KRAS mutation impacts outcomes. In this study, we investigated the clinical and molecular features of PIK3CA mutations with and without KRAS mutations, and examined the impact of KRAS mutations on clinical outcomes of patients with PIK3CA-mutated CRC. Methods: We reviewed 147 patients at our institution diagnosed with CRC cancer whose tumors harbored a pathogenic PIK3CA mutation with or without KRAS mutation. Demographic and clinical features, including age, gender, race, tumor location, and site of initial metastasis, were collected from review of electronic medical records. Fisher’s exact test was used to investigate the association between the aforementioned features and KRAS mutation status. For survival analysis, we included patients with metastatic disease (de novo or metachronous) using Kaplan-Meier survival and univariate Cox regression analysis, with statistical significance defined as a p-value threshold of <0.05. Results: Within our initial cohort, 55.1% (N=81) presented with local disease (stage 1-3), while 43% (N=63) had de novo metastatic disease at time of diagnosis. The median age was 66.3 years. An equal proportion of patients were male (50%) and female (50%), and predominantly identified as White (83%). Most primary tumors were colon (78%) and left-sided (52%). Site of initial metastasis tended to be liver (49%), rather than non-liver (32%), and a minority had both sites (19%) at time of stage IV diagnosis. There was no significant difference in demographic features between KRAS (N=94) and non-KRAS (N=53) patients, aside from stage at diagnosis. In our analysis of PIK3CA patients with metastatic disease (N=102), patients with concurrent KRAS mutation (N=69) had a median progression-free survival of 17.7 months (95% CI: 14.2-23.3, p = 0.01). Without KRAS, the median progression-free survival was 32 months (95% CI: 24.7 – not reached). Presence of KRAS mutation conferred a hazard ratio (HR) of 2.59 (95% CI: 1.22-5.48, p = 0.013), indicating worse progression-free survival. With the exclusion of MSI-high disease, the overall survival of patients with concurrent KRAS mutation was 33.7 months and it was not reached for those without KRAS mutation (95% CI: 24.7-56.6, p = 0.049) with a HR 2.23 (95% CI: 0.98-5.08, p = 0.055). Specific locations of KRAS mutations (exon 2 vs non-exon 2) had no impact on clinical outcomes. Conclusions: KRAS mutations confers poor prognosis for patients with PIK3CA mutated CRC, and highlights biological and clinical differences between KRAS Wild type- PIK3CA mutated CRC and those with a KRAS mutation.