No abstract

836 Background: Treatment of metastatic colorectal cancer (mCRC) has greatly advanced over the past decade, based on data from randomized controlled trials (RCTs). This raises the question whether results of RCTs, performed on selected patients (pts), do reflect outcomes in real-life practice. The aim of this study was to summarize our experience in the treatment of mCRC and compare it to data reported in RCTs. Methods: A retrospective single-institution study on consecutive mCRC pts treated with first-line bevacizumab-containing regimens in our institute between 2006 and 2014. Results: The study included 300 pts, of whom 54% were males. Median age was 67 years (range 28-90), 26% aged ≥ 75 years. ECOG performance status was ≤1 in 93%. The primary tumor site was right colon in 37%, left colon in 40%, rectal in 23% and 1% of pts had synchronous tumors. RAS status was available in 60%, of whom 55% had wild-type alleles. 46% of pts had a single metastatic site, including 27% with liver-limited disease, and 54% had multiple metastatic sites. Irinotecan-based chemotherapy was used in 66%, oxaliplatin-based chemotherapy in 29% and flouropyrimidine monotherapy in 5%. Curative metastasectomy during 1st line treatment was performed in 29%. Grade ≥3 hematological and non-hematological toxicities were reported in 24% and 38% of pts, respectively. Second and third line treatments were administered to 75% and 66% of pts, respectively; 73% of pts received both irinotecan and oxaliplatin through their treatment course and 76% of those with wild-type RAS were treated with anti-EGFR therapy. Overall response rate and disease control rate were 69% and 89%, respectively. Median progression-free survival (PFS) and overall survival (OS) were 17 and 28 months, respectively. In a sub-group analysis on "RCT-like population", excluding pts ≥ 75 years, ECOG PS ≥ 3 and/or mutated/unknown RAS status, median PFS and OS were 15 and 29 months, respectively. Conclusions: The results of this study suggest that, if adhered to international clinical guidelines, outcomes reported in RCTs are indeed reproducible in routine clinical practice in unselected real-life pts. Additional data, with more pts and longer follow-up, will be presented.