3060 Background: Optimized phase 1 (ph1) clinical trials may increase the likelihood of successful oncology drug development. We developed a FDA knowledge management (KM) database that allows comprehensive assessment of ph1 trial features. Here we pilot whether the database can catalogue diversity in early phase studies of the PI3K inhibitor drug class. Methods: We evaluated the development programs of 6 PI3K inhibitors, including ph1 protocols, annual reports, and investigator brochures submitted to the FDA under the investigational new drug (IND). Key features of PI3K inhibitor ph1 trials were summarized in the KM database. Results: Of the 6 PI3K inhibitor ph1 trials surveyed, three agents (50%) targeted PI3K alone and 3 (50%) were dual PI3K and mTOR inhibitors. Four (66%) trials used a traditional 3+3 design, while two (33%) adopted a bayesian continual reassessment model. MTD was the primary endpoint of all trials. Three trials (50%) adopted an enrichment strategy for the expansion cohort, which varied from molecular enrichment (somatic alteration in PI3K, PTEN or EGFR) or clinical enrichment (Cowden syndrome). Pharmacodynamic (PD) assays included surrogate tissue (assessment of molecular changes in PBMCs, skin, and hair) and primary tissue (FDG-PET, flow cytometry of CLL, and tumor biopsy). Dose-limiting toxicities included hyperglycemia, anorexia, gastrointestinal, rash, mood alteration, uveitis, thrombocytopenia, elevated ALT, low phosphate, and ECG alteration. Two trials (33%) reported ≥ grade 3 hyperglycemia. Conclusions: FDA review of PI3K ph1 studies reveal a variety of designs, enrichment strategies based on molecular and clinical features, PD assays, tumor types, and toxicities. The KM database will assist in determining which features are important to successful development. Preliminary analysis of these trials (as well as prior experience with other targeted agents) indicate that early selection of an enriched population most likely to benefit, development of validated biomarker assays prior to pivotal trials, adoption of randomized ph2 designs, and exploration of combination strategies will accelerate the development programs and increase the likelihood of success.