The research work summarized in this thesis is largely focused on the role of KRAS and BRAF mutation status in determining prognosis and guiding management among patients with resectable CRLM. The decision to focus our investigation on these two biomarkers reflects both their central importance in colorectal carcinogenesis and the availability of relevant data which are collected as part of routine practice in most institutions. The publications presented in this work follow a well-defined continuum that traces the evolution of a number of interconnected ideas on the impact and possible role of available genetic data in the clinical management of patients with CRLM. This intellectual journey begun in 2015 with our first report confirming the prognostic role of KRAS mutation status in CRLM. Subsequently, we proceeded to demystify the impact of BRAF mutations on outcomes which at the time were considered synonymous with poor survival. While we confirmed that BRAF mutation status was indeed a potent predictor of poor prognosis, we were also able to identify distinct subgroups with favorable outcomes (i.e., patients with non-V600E mutations) and ultimately dispel the idea that these patients do not benefit from operative management. Despite these discoveries, when we sought to improve prognostication among patients with CRLM by introducing genetic biomarkers into contemporary risk scores, we found ourselves obligated to employ KRAS, rather than BRAF, due to the much higher incidence of KRAS mutation in CRLM cohorts. The result of this effort was the GAME score, which has gradually emerged as the most robust and accurate of a series of “hybrid” prognostic models that combine genetic and clinicopathologic variables. More so than its successes, the limitations of the GAME score and its predecessors proved of vital importance to future research endeavors by identifying persistent deficits in prognostication among patients with CRLM. One such example was the realization that contemporary survival models offered a static picture of projected outcomes at the time of surgery, failing to reflect clinical developments in a dynamic fashion. Conditional survival analysis helped address this limitation, confirming the prognostic discrepancies between patients with different BRAF codon mutations and highlighting the long-term importance of traditional variables such as surgical margin. The idea that tumor biology could impact tumor growth patterns and thus dictate surgical technique most likely to result in disease control led us to identify the possible utility of anatomical hepatectomy among patients with KRAS mutations. While definitive proof of efficacy will need to await dedicated clinical trials, this work initiated a debate on the potential of “precision surgery,” which remains active to this date. Lastly, having largely explored the prognostic implications of KRAS mutation status among patients with CRLM, we employed this variable to help reconcile disparate reports on the effect of tumor laterality on outcomes, paving the way for the utilization of this powerful prognostic factor in future risk scores.
