KRAS aKtive, an Italian network for assessment of KRAS mutations in colorectal cancer patients: Results on 7,432 cases.

e14042 Background: The KRAS aKtive program was started on March 2009, promoted by the Italian Association of Medical Oncology (AIOM) and the Italian Society of Surgical Pathology and Cytopathologyy (SIAPEC) to support the activity of oncologists and pathologists involved in the management of metastatic colorectal cancer patients who need the assessment of the mutational status of the KRAS gene. Methods: The program was specifically devised to facilitate the exchange of biologic material, clinicopathological data and diagnostic reports within a network of oncologist, pathologists and pathology/molecular biology reference laboratories throughout Italy, connected through the site www.kras-aKtive.it. KRAS mutation analysis was performed by Sanger sequencing (SS), real time PCR or other techniques, including pyrosequencing and hybridization strip assays. Data were collected in a common database. Results: The KRAS aKtive program has involved 478 oncologists, 144 pathologists, and 24 reference laboratories. A total of 7,432 KRAS mutations analyses were performed. The tests were informative in 7,265 cases (98%). The vast majority of tests (5,626 cases, 77.4%) were conducted by SS. In 529 (7.3%) cases a real-time PCR assay was used, other detection techniques were used in 1,110 (15.3%) cases. KRAS mutations at codons 12-13 were detected in 2,874 cases (39,6%). The frequency of mutations detected by real-time PCR or other techniques (45%, and 43%, respectively) was significantly higher (p=0.002, and p=0.008%, respectively) than that observed by SS (38%). The percentage of cases evaluated by non-SS-based methods has increased during the first three years of the program. Conclusions: The results of this large survey allow an accurate estimation of the actual prevalence of KRAS mutations and their types in caucasian colorectal cancer patients. Our data indicate that the frequency of mutations detected by non-SS-based methods is higher than that obtained by SS.