Early-phase clinical trial referral and outcomes in breast cancer: A prospective screening cohort and a real-world Phase I program experience at Gustave Roussy.

Background: Early-phase clinical trials are increasingly relevant for metastatic breast cancer, but referral timing and benefit remain insufficiently documented. Methods: We analyzed two complementary cohorts. The first comprised 231 consecutive metastatic breast cancer patients prospectively screened for early-phase trial consideration at Gustave Roussy. The second included all breast cancer patients who initiated treatment in a phase I trial (C1D1 cohort) since 2009. Baseline characteristics, treatment history, referral decisions, RECIST outcomes, progression-free survival (PFS), and overall survival (OS) were collected. Results: In the prospective cohort (n=231), most patients had HR+/HER2- (54%) or TNBC (40%) disease. Only 18% underwent molecular profiling before referral review. Overall, 39% were referred to a phase-I trial, while the main reasons for non-referral were absence of available trial slots (37.3%), excessive prior treatment lines (36.4%), lack of measurable disease (7.4%) and brain metastases (6.2%). Patients entering phase I screening had received fewer prior therapies (median 3 vs 6) than those excluded, highlighting that late referral frequently precluded eligibility.In the C1D1 cohort (n=343), median age at treatment start was 52 years and patients had received a median of 3 prior lines of therapy. TNBC accounted for 48% of treated cases, followed by HR+/HER2- (45%) and HER2+ (7%). Most patients received targeted therapy (66%), immunotherapy (20%) or intratumoral therapy (8.1%). Median PFS and OS were 3.2 and 11.5 months. HER2-positive disease achieved the longest PFS (4.24 months), followed by HR+/HER2- (3.19 months) and TNBC (2.86 months). The overall response rate was 13.4% and the clinical benefit rate was 52.8%. The highest responses were observed in HER2-positive tumors (31.8%), whereas TNBC demonstrated the lowest (~7%). Targeted therapies achieved the best outcomes (median PFS 3.7 months), while intratumoral approaches showed limited benefit (2.1 months). Conclusions: Most patients in the prospective cohort were referred late and became ineligible due to protocol-defined treatment history or disease burden. Early molecular profiling and proactive referral are essential to preserve eligibility. Patients who receive phase I treatment derive meaningful benefit, particularly with targeted approaches, supporting earlier integration of phase I programs in the breast cancer therapeutic sequence.