Abstract Introduction: Despite a decade of metastatic colorectal cancer (mCRC) therapeutic innovations, a comprehensive synthesis of patient outcome and risk-benefit assessment of phase 1/2 trials, extending beyond single-center studies, is missing. The aim of this meta-analysis is to assess efficacy, safety, and trends over time for phase 1 and 2 mCRC trials by examining clinical benefit rate (CBR), overall response rate (ORR), grade 3 or higher adverse events (AE), and discontinuation due to AE. Methods: The protocol and search were preregistered on PROSPERO. We searched PubMed and Embase for publications of phase 1/2 trials between 2010-2022. Reports were screened for inclusion criteria, followed by data-extraction by two researchers. Trials reporting on non-standard therapies for treatment-refractory mCRC or trials reporting specific mCRC efficacy data were included. CBR was defined as the proportion of patients with complete/partial response or stable disease ≥14 weeks. We used logit transformations for pooled effect analysis via a random-effects model, assigning study weights by the inverse variance method. Subgroup analyses were performed for publication year, preselection on molecular traits, and drug class: antiangiogenics, chemotherapy, immunomodulator, targeted therapy, or vaccine. Data are presented with 95% confidence intervals (CI); reported differences had non-overlapping CIs. Results: The search strategy yielded 4175 unique reports, from which we extracted data for 279 trials. Most trials were phase 1 (63%). ORR was limited to 6% [6-7] in both phase 1 and 2, CBR was higher in phase 2 than in phase 1 trials (36% vs 27%). Trials testing antiangiogenic drugs had the highest CBR (47%) and immunomodulators the highest ORR (8%). Trials with patients preselected on molecular traits had higher CBR compared to other trials (38% vs 28%). We observed an increase in efficacy, especially in phase 2 trials, with a CBR rising from 22% to 42% between trials published in 2010-2012 and 2019-2021. Compared with 2010-2012, trials in 2019-2021 were more often testing immunomodulators (23% vs 4%), performed on molecular preselected populations (38% vs 4%), and including younger patients (median age <60 in 66% vs 44%). The occurrence of grade 3+ AE was 33% [28-38] in both phase 1 and 2. AE led to discontinuation in 12% of phase 1 trial patients, and 7% of phase 2 patients. Grade 3+ AE occurred in 4% of vaccine trial patients, in contrast to 41% in trials on targeted therapies. Occurrence of grade 3+ AE did not change over time. Conclusion: We found limited ORR and CBR in early phase trials conducted in 2010-2021 for mCRC patients, while noting severe AEs in many patients. However, during this period, clinical benefit rate increased for phase 2 trials, possibly due to evolving study strategies that entail testing new treatment types in a younger, fitter, and stricter preselected study population based on molecular traits. Citation Format: Maarten A. Huismans, Lidwien P. Smabers, Niels N. van Nieuwenhuijzen, Monique C. Minnema, Miriam Koopman, Hugo J. Snippert, Anne M. May, Jeanine M. Roodhart. Meta-analysis of efficacy and safety in early-phase clinical trials for refractory colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2386.